Dear {{ first_name | reader }},

I'm Belgian, and I won't pretend that doesn't colour this week's edition. Belgium beat the United States 4–1 in the World Cup last 16, and I felt the particular pride of watching your team win with something to prove.

To the American half of this list: I'll try to be a gracious winner. I will not succeed.

Here's why you're reading this… FIFA reversed a suspension the day before kickoff, then barely survived a Belgian appeal hours before it. Belgium won anyway and trolled FIFA and Trump on Instagram for it.

Months earlier, Iran had already built a whole propaganda genre on the identical script. A leader leans on a technical process. The process bends. Someone turns the bending into a symbol.

If you work in crisis, risk or emergency communication, that pattern (not the football) is what this edition is actually about.

Enjoy!

The case file: what actually happened

Folarin Balogun picked up a straight red card in the USA's 2–0 win over Bosnia and Herzegovina the previous Wednesday. That's meant to trigger an automatic one-match ban, no discretion involved.

On the Sunday before the Monday night tie against Belgium, FIFA invoked Article 27 of its own disciplinary code to suspend the implementation of that ban for a one-year probationary period — not a dismissal, technically, but the practical effect was the same. Balogun was free to play.

Belgium wasn't done. The Belgian federation was granted an appeal, and the dispute over that ban ran live right up to matchday — FIFA rejected the appeal only hours before kickoff in Seattle. UEFA called the reversal "unprecedented, incomprehensible and unjustifiable" and said the decision had "crossed a red line".

President Trump had phoned Infantino to press for the review, then publicly thanked FIFA on social media for reversing what he called a great injustice.

Belgium won 4–1 anyway. Midfielder Nicolas Raskin said afterwards that the result felt like a measure of justice. The team's Instagram account posted a photo of Lukaku celebrating Belgium's fourth goal, silencing the crowd, captioned with two words: "Overturn this" – aimed as much at FIFA and Trump as at the scoreline.

Why this became a case study in politicisation risk

For USA head coach Mauricio Pochettino, the story was supposed to be about tactics and squad management, not politics. Trump's intervention undid that plan before the match even started. Distraction, not preparation, is what followed instead.

Commentary on the game repeatedly pointed to exactly that: presidential lobbying introducing the kind of chaos Pochettino had been trying to keep out of his own dressing room.

The psychological pressure compounded when the US defence collapsed. The team was eliminated from what was meant to be their home tournament. Easy, after that, to blame hubris and political favour instead of the football. Fair or not, that's the narrative that stuck.

This is the risk-communication lesson buried in the football commentary. Political leaders who intervene in technical or adjudicative processes – refereeing panels, safety regulators, and ethics committees – can win the immediate battle and still lose control of the longer story.

The intervention itself becomes the headline. Performance failures that follow get read through that lens, whether or not they're actually connected.

Sport has always been soft power — what's new is the speed

None of this is new in the abstract. Academic literature has treated sport as an instrument of soft power for decades – hosting rights, sportswashing, boycotts, and national identity performance through a team shirt.

Mark Galeotti calls it the "weaponisation of everything". Single matches, not just Olympic bids or World Cup hosting decisions, now function as stages for national signalling.

What's changed is the speed and the stitching.

The propaganda genre you should already be tracking

Separately from the World Cup, Iranian state media broadcast the first video in what became a whole genre — an AI-generated short called "Narrative of Victory", released on 10 March 2026. It went viral fast.

State-linked accounts, including outlets aligned with the IRGC, spent the months since circulating Lego-style animations depicting Trump and allied leaders as reckless, cartoonish villains. In the original video, a Lego Trump — egged on by a Lego Netanyahu and a devil figure — fires a missile that hits a girls' school, rendered as a pair of small shoes left in the rubble.

A grieving Iranian soldier then launches retaliatory strikes of his own. A dozen-plus follow-up videos expanded the format, some set to rap, and racked up hundreds of millions of views between them.

Explosive Media, the account behind that first video, wasn't working alone. A parallel creator called PersiaBoi — real name Mahdi Hemmat, an Iranian-American — built his own following with Lego-style rap videos under the banner PersiaBoi Studios. He's said the strike on that same girls' school was what pushed him to start.

His output has tracked the war almost in real time since: "Uranium Heist in the Dead of Night" mocked a US pilot rescue mission; "Fake Jesus" turned Trump's deleted AI self-portrait into a diss track within hours of the post going up. Millions watched. Some clips landed on state media. Others didn't need to — they'd already gone viral on their own.

PersiaBoi’s LEGO clip opening scene.

Whether Hemmat is an independent artist working out real anger or a node in a wider state messaging operation is genuinely contested — Explosive Media has made the same claim of independence while acknowledging Iranian state media as a paying customer.

That ambiguity isn't a flaw in the propaganda. It's the mechanism. A viewer trusts a lone creator with a grievance far more readily than a government broadcaster, and the format is built to make that distinction impossible to check.

Analysts have described this genre as "slopaganda" — cheap, fast, highly shareable content built for volume rather than nuance, designed to overwhelm careful rebuttal with sheer repetition. The aesthetic does real work: children's building blocks lower a viewer's guard in a way a state television broadcast never could.

The Americans play this game too, just badly. The White House has pushed out its own AI-generated content since the war began – bragging at one point about more than two billion impressions – but outside analysts have judged it heavier on stock explosions than on story, and it hasn't travelled the way Iran's has.

That asymmetry is the actual finding here, not who's right in the underlying conflict. Effectiveness, not righteousness, is the variable a communications professional should be tracking.

Iran has also used its own World Cup fixtures for political signalling, quite apart from the Lego videos. Officials described the national team as facing unfair treatment over US visa restrictions and travel conditions between matches — head coach Amir Ghalenoei called Iran "the most oppressed team in the whole World Cup".

After a goalless draw with Belgium in Los Angeles, the Iranian squad left a handwritten note in the dressing room calling for peace and mutual respect, tagged with the hashtag #168 – a reference to the same Minab school strike that seeded the Lego propaganda genre in the first place. Two entirely separate messages, one shared wound.

None of this explicitly cites the Balogun reversal or Belgium's win. It doesn't need to. The narrative architecture is identical: a political leader meddles, the meddling backfires or looks unjust, and the backfire gets repackaged as proof of American overreach.

Splice the football clip, the Trump–Infantino calls, and a Lego animation together. The implicit message writes itself: wherever Trump intervenes, he loses, and someone else gets to look righteous.

The real complication

I want to push back on my own framing before you do. The "politics always backfires" narrative is satisfying, but it's still a narrative – built after the fact by people with an interest in telling it that way.

Sometimes political pressure on a technical body works exactly as intended, and nobody notices because there's no dramatic loss to turn into a story. We're seeing the cases where the intervention was visible and the outcome was bad. That's a selection effect, not a law of nature.

The Iran connection deserves the same precision. There's no confirmed direct link between the Lego propaganda campaign and the USA–Belgium match itself. The overlap is thematic, not coordinated.

Treating every anti-US meme as part of one orchestrated campaign risks the same crude pattern-matching we'd criticise in a client who sees conspiracy in every negative headline. The honest version of this piece is that two separate narrative streams share a script — not that one caused the other.

What this means for your monitoring and your messaging

Three things I'd actually act on if I ran communications for an organisation with a technical or adjudicative function right now.

1) Build independence messaging before you need it.

Get your disciplinary panels, safety regulators or ethics committees communicating their criteria and safeguards now — not defending their integrity for the first time in the middle of a crisis.

2) Widen your horizon scanning past hard news, too.

Sport, entertainment and meme ecosystems shape public sentiment about legitimacy and fairness long before that sentiment surfaces in a genuine security or diplomatic incident. Stop at press and policy, and you're always two weeks late.

3) And plan for a remix, not just for the statement.

Assume whatever you publish gets turned into a symbol, a slogan, or a piece of visual shorthand faster than you can issue a follow-up. Which lines and images are most likely to get lifted out of context? Work that out before you publish, not after.

The separation principle is old advice: keep politics out of technical decisions wherever you can. What Belgium–USA and the Lego memes both demonstrate is what happens once you don't – and how little control you get back once the story starts writing itself.

What to take back to your desk

I enjoyed Belgium's win more than professional detachment probably allows – sorry again, American readers; still not sorry. 😅

But the part of this that will stay in my training materials isn't the scoreline. It's how fast a disciplinary reversal, a phone call, and a piece of children's-toy animation ended up telling the same story about power, legitimacy, and who gets to look righteous when it all goes wrong.

Worth Attending + A Gift For You

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

SPONSOR

200+ Claude Prompts Top Professionals Actually Use at Work

Claude can be your analyst, editor, and strategist.
But most professionals are using it to fix grammar.

These 200+ Claude prompts take it from grammar tool to your most powerful AI work assistant.

Sign up for Superhuman AI and get:

• 200+ ready-to-use Claude prompts to get real work done in minutes — researched, tested, and used by professionals at Google, Microsoft, and NASA

• Superhuman AI newsletter (4 min daily) so you keep learning new AI tools and skills to stay ahead in your career — the prompts are just the beginning

Claim your free prompts

Dear {{ first_name | reader }},

This week's article leans on a study from IBM's Institute for Business Value, and a disclosure is worth making before you read another word: I spent ten years inside IBM earlier in my career as Public Relations Manager.

I still read their research through that lens. Not a conflict of interest but a bias worth naming rather than leaving you to guess at it.

That history aside, the numbers stopped me because they describe a crisis type most of us haven't built a plan for. Not the kind where AI generates a bad citation or a convincing deepfake. The kind where an organisation discovers, mid-disruption, that it never mapped what it was standing on in the first place.

So this week: three ways AI can sit at the origin of a crisis, why only one of them looks like the crises we're used to planning for, and what a global study of 1,000 executives says about how exposed most organisations actually are.

Enjoy!

Most of your AI crisis planning — if you have any yet — is built around the same assumption.

The AI does something wrong. It hallucinates a citation, fabricates a statistic, and generates a deepfake, and the crisis follows from that failure. That assumption isn't wrong. It's just incomplete.

IBM's Institute for Business Value published a study this month surveying 1,000 senior executives across 16 countries and 17 industries. The headline number: 91% of respondents say they don't fully understand their own organisation's dependencies across AI vendors, models, and infrastructure.

Not "AI failed us". They don't know what they're standing on in the first place. That's a different origin point entirely. And it's the one most crisis communication frameworks – mine included, until now – haven't named.

The three origins, not one

When AI sits at the root of a crisis, it tends to arrive by one of three distinct routes. I've been building a framework around this distinction, and the IBM data is the clearest evidentiary anchor I've seen for the third and least discussed of the three.

Origin one: AI causes the crisis directly. 

A hallucinated report goes to a client. A deepfake fabricates a statement from an executive who never made it. An autonomous system makes a consequential decision on bad data. The AI did something, and that something is the crisis.

Origin two: AI complicates an existing crisis. 

The crisis started elsewhere — a product recall, a data breach, a natural disaster — and AI accelerates or distorts the response. Misinformation spreads at machine speed. Algorithmic amplification picks the wrong narrative and runs with it. AI didn't start the fire. It's pouring fuel on one that was already lit.

Origin three: structural dependency exposure. 

No hallucination. No deepfake. No bad actor. A vendor deprecates a model. A price hike without warning. A jurisdiction changes its data residency rules overnight. The organisation discovers, mid-disruption, that it never mapped what it actually depends on — and the discovery itself is the crisis.

Origin three is the one your response plan almost certainly doesn't cover, because it doesn't look like a crisis until it already is one.

What the numbers actually say

The IBM figures are worth sitting with, because they describe a structural condition, not an isolated incident.

Seventy-one per cent of respondents say switching their primary AI vendor or model would be difficult. That's lock-in. Sixty-eight per cent say meeting data residency and sovereignty requirements across geographies is challenging. That's regulatory exposure baked into the infrastructure layer, not the content layer.

Then the disruption data. Surveyed leaders report an average of six AI-related disruptions over the past two years, driven largely by vendor services. And 81% say a seven-day vendor outage would still cause severe or critical disruption — effectively halting operations. Six disruptions in two years, and no resilience built for the seventh. That's not misfortune. That's a system with no shock absorber.

There's a control-maturity split underneath all of it.

Organisations with the most advanced AI control capabilities protect 55% more operating profit from AI-driven disruptions than their peers. Only 7% of surveyed organisations operate at that level. Ninety-three per cent are exposed to some degree, and most of them can't quantify by how much – because the 91% figure and the 7% figure describe the same population from two different angles.

Why this origin type gets missed

Crisis communication doctrine is built for events with a clear trigger and a visible start point. A recall has a date. A breach has a discovery moment. A deepfake has a public timestamp.

Structural dependency exposure doesn't announce itself. It sits underneath operations for months or years, invisible, until a vendor decision — not yours — turns it into an active crisis in a single afternoon.

There's no bad actor to name and no failure to point at. The organisation's own unreadiness is the story, and that's a harder narrative to run a response around than "we made an error, and we're fixing it."

It's also, honestly, harder to sell internally as a risk worth planning for. "Our AI hallucinated" gets the budget approved for governance. "We don't fully understand our own vendor dependencies" sounds like an IT hygiene problem until the seven-day outage hits, and it's suddenly the only problem in the building.

The complication

I'll be straight about where this framework is still thin. The IBM study measures perception and structural exposure – it doesn't yet give us a clean causal chain from "we didn't understand our dependencies" to "here's the crisis communication playbook that followed."

That case-level evidence is still being written, largely because organisations that experience this kind of crisis tend to describe it afterwards as an operational or vendor-management failure, not a communication one.

The communication dimension — how you explain to clients, regulators, and staff that your outage happened because you didn't map your own stack — is underexamined territory. That's partly why it's worth naming now, before the case studies pile up and the doctrine gets written by people who missed the pattern.

What to actually do with this

Three moves, in order of how fast you can act on them.

First, separate your AI crisis planning by origin type. A hallucination playbook and a structural-dependency playbook require different stakeholders in the room, different disclosure postures, and different timelines. Don't run them off the same template.

Second, ask your own version of the 91% question. Not "is our AI accurate?" — "do we know what we'd lose, and for how long, if any single AI vendor became unavailable tomorrow?" Most organisations have never asked this out loud. If you can't answer it in under a day, you've found your gap.

Third, treat sovereignty as a communication asset, not just an IT one. If you can say, credibly, that your organisation is designed for vendor and model flexibility, that's a legitimacy point you can make to clients and regulators before a disruption — not just a technical fact you'd cite defensively after one.

I've built my own practice around a version of this. A sovereign stack running across an EU inference layer and local models, reserved specifically for client-sensitive work, with a separate layer for everything else.

It's a small-scale version of exactly what IBM is describing at enterprise scale. The logic holds at both ends: know what you depend on, or find out the hard way when someone else's decision becomes your crisis.

Frequently asked questions

What are the three ways AI can be the origin of a crisis? AI can cause a crisis directly (a hallucinated report, a deepfake, a bad autonomous decision), complicate an existing crisis by accelerating or distorting the response to it, or expose an organisation through structural dependency — where a vendor, model, or infrastructure change triggers a crisis with no AI failure involved at all.

Why does structural dependency exposure count as an AI-origin crisis if the AI didn't do anything wrong? Because the crisis originates from the organisation's relationship to the AI system, not from the AI's output. A vendor deprecating a model, changing pricing, or restricting usage can halt operations just as effectively as a hallucination – the difference is that no one made an error, which makes it harder to communicate and harder to plan for.

What does the IBM study actually show? IBM's Institute for Business Value surveyed 1,000 senior executives and found that 91% don't fully understand their organisation's AI dependencies across vendors, models, and infrastructure, that 71% would find switching their primary AI vendor difficult, and that organisations with advanced AI control capabilities protect 55% more operating profit from AI-driven disruptions than the 93% who don't operate at that level.

Where does this framework fall short right now? The evidence base is structural, not case-level — IBM's data shows the exposure exists, but the communication playbook for how organisations actually explain a dependency-driven crisis to clients and regulators is still thin, largely because these incidents tend to get classified as vendor-management failures rather than communication events.

How is this different from standard AI misinformation or deepfake risk planning? Misinformation and deepfake risk sit in the first and second origin categories – AI causing or amplifying a crisis through its output. Structural dependency exposure requires no AI output at all to go wrong. It requires only that the organisation never mapped what it depends on, which means the mitigation is an audit exercise, not a content governance one.

What should a crisis communication team do with this framework now? Build separate response playbooks for each origin type rather than one generic "AI crisis" plan, run an honest internal audit of what would happen if a single AI vendor became unavailable for seven days, and start treating demonstrated AI sovereignty as a proactive credibility asset with clients and regulators rather than a defensive talking point held in reserve.

Worth Attending + A Gift For You

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

SPONSOR

Your membership list is full. Your calendar isn't.

You have hundreds of members needing tune-ups twice a year.

Scheduling them means multiple calls and weeks of manual outreach that never fully gets done.

The ones who don't get reached fall through the cracks. Those are members and recurring revenue quietly walking out the door.

Podium's AI Membership Coordinator texts every member, books the appointment, follows up automatically, and handles renewals.

It works your entire list proactively so your slow season fills before it starts.

No manual outreach. No dropped members. No empty shoulder season.

Make the most of your membership list.

Fill My Seasonal Calendar

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

Dear {{ first_name | reader }},

This week I want to take you somewhere most crisis plans never look: the network the message travels through, not the message itself.

A few days ago I read Christian Johnson's account of cascading failure on a power grid, and one detail in it explained why some crises stay contained and others run away, almost regardless of how good the messaging was.

I've turned that logic into something you can actually use, and a (fictional) water utility's “brown-tap scare during a heatwave” scenario shows exactly how it could work.

Here's what's in this edition.

Enjoy!

A cascade is a cascade — up to a point

It is 43 degrees, the third day of the heatwave, and a pump station has just tripped. Pressure falls across the northern districts. Within twenty minutes, someone films the brown water coming out of their kitchen tap and posts it:

The water is fine. The discolouration is harmless sediment, shaken loose by the pressure drop. But it doesn't look fine, and by mid-afternoon "don't drink the tap water" is climbing the regional trends — during a heatwave, where not drinking water is the one thing that will actually hurt you.

This is the part most crisis plans miss. The utility's problem was never the message. It was the network the message landed in.

We were trained to treat a crisis as a discrete event — a thing that arrives, gets handled, resolves, and lets attention move on. A polycrisis doesn't behave like that.

Crises arrive coupled now. A physical failure seeds an information failure; the information failure changes behaviour; the behaviour makes the physical danger worse. The pump, the rumour and the heat were three problems wearing each other's clothes.

Back to Christian's account of modelling cascading failure in a national power grid.

With protections switched off, a single fault spread across the network and took down huge sections at once. With protections switched on, the same fault was contained before it became a disaster. Same fault. Different outcome. The difference was design.

One phrase from that account stuck with me: a cascade is a cascade. The mathematics, the engineer argued, doesn't much care whether it's modelling a power grid, a market, a supply chain or a country.

That's true — and it's also where you have to be careful, because the point at which it stops being true matters.

The equations don't transfer wholesale, and anyone who tells you a tweet and a transformer obey the same law is selling something.

What does transfer is the shape of failure. Every cascade — electrical, financial, viral, reputational — has a line. Below it, a local shock fizzles out. Above it, each affected node hands the failure to more than one other, and the whole thing runs away.

The craft of containment is keeping your communication network on the safe side of that line and knowing at all times how close to it you are.

Why this changes the job

Most crisis communication optimises the message: the wording, the holding statement, and the tone. That's necessary work.

But in a polycrisis, polishing the message while ignoring the network is tuning the instrument while the hall fills with smoke. What decides whether your crisis stays contained is the network itself — how fast it's moving, how much trust it's carrying, and where it's brittle.

Two things follow from that, and most practitioners already half-know both.

Trust isn't something you summon during the crisis. It's a reserve.

You either deposited it beforehand – by being credible, fair and legitimate when nobody was watching – or you didn't, and no holding statement will conjure it at three o'clock on the worst day. In the network, trust is carrying capacity. It's what lets a message actually move across a relationship instead of bouncing off it.

And there is no "general public" to reach. There never was.

The utility's problem wasn't "the public". It was 180,000 people in specific districts who, after a billing dispute the year before, didn't believe a word the company said. You don't reach those people by broadcasting louder on the channel they already distrust. You reach them through the people they do trust — which you can only do if you mapped those people before the crisis arrived.

The work is done in the calm

So why didn't that cascade run away?

Not because of quick thinking on the day. Because of three dull decisions made months earlier.

The utility had pre-written a 200-word explainer for exactly this scenario — harmless discolouration after a pressure drop — so it could acknowledge within twenty minutes instead of going silent while it worked out what to say.

It had a current list of community leaders in the northern districts, so "the water is safe, keep drinking in this heat" arrived through trusted local voices rather than the company's own distrusted feed. And it had found those gaps the only way anyone ever does: by running a rehearsal built to break, three months earlier, and fixing what broke.

None of it was heroic. The most valuable thing the team owned that day was a short note nobody had thought they would need.

What the engineering actually maps to

I've spent the past few months turning that grid logic into something a communication team can run, not just admire from a distance. The protections that stop a fault from cascading across a power network have direct equivalents at the level of attention and trust, and naming them is what makes the analogy useful instead of just evocative.

• A circuit breaker becomes a velocity throttle – a phased release that stops an unverified claim accelerating past the speed at which anyone can check it.

• Islanding becomes channel isolation — quarantining a compromised platform so its failure doesn't pull the rest of your channels down with it.

• Load shedding becomes message triage — cutting volume to match what an exhausted audience can actually absorb before the signal turns to noise.

• A hidden-failure audit becomes a plan rehearsal — the only way to surface the stale contact list or the single-person credential before a crisis surfaces it for you.

• Relay coordination becomes a cross-channel discipline, so units don't contradict each other into a confusion the original event never caused.

• And redundant paths become a trusted-messenger network — the line that keeps carrying your message when the official channel is down, overwhelmed or simply not believed.

None of these works alone. A utility with excellent message discipline but no rehearsal will still be undone by a hidden failure it never tested for. One with strong messenger redundancy but no velocity control will watch an unverified claim outrun every correction it issues.

Containment is a property of the full set working together – which is exactly why it has to be designed before the pump trips, not assembled by instinct while it's failing.

The honest limitation

This framework borrows the grid's protection philosophy and its method for proving value. It doesn't borrow the grid's physics, and it shouldn't.

The moment a model imports conserved-flow mathematics into a system where the thing flowing actually multiplies, the model stops describing reality and starts performing rigour it hasn't earned.

The discipline here isn't pretending communication networks and power grids are identical. It's being precise about exactly where they rhyme and exactly where they don't and building only on the part that holds.

There's a second limitation worth naming plainly. None of this is a forecasting tool.

Knowing where your network sits relative to the containment line tells you whether you're operating with margin or sitting one shock away from losing it — it doesn't tell you which shock arrives next, or when. The value is in the readiness it buys you, not in prediction it can't honestly offer.

What to do with this tomorrow?

You don't need the full architecture to start. Pick one protection from the six above — the velocity throttle, the channel isolation, the trusted-messenger network, or any of them — and ask whether you actually have it or whether you've only assumed it.

Most organisations discover they have message discipline and nothing else: no velocity throttle, no rehearsed channel isolation, no trusted-messenger list that's been updated since the last reorganisation.

Then ask the harder question: when did you last rehearse hard enough to find a failure that was lying dormant? Not a tabletop that confirms what you already believed — a rehearsal built to break something.

If you can't point to a specific gap your last drill surfaced and fixed, you don't have a tested plan. You have a hypothesis with a nice cover page.

The utility in this story didn't get lucky. It had spent months making itself boring in exactly the right places so that on the one day it mattered, nothing dramatic needed to happen.

Where this is going

I've been building this out into a full method — how you map the network, read how loaded it is, and arm the protections before the shock lands.

It sits as the network-resilience layer inside the wider Communication Nexus Framework, and it draws on the GATE model as its analytical engine once a crisis is live.

I'll come back to it properly in a future edition, with the full sequence and the maturity model that tells you exactly where your organisation sits today.

That decision isn't made in the crisis. It's made today.

Philippe

FAQ

What is cascade containment in crisis communication? Cascade containment is the practice of managing a crisis at the level of the network it travels through, rather than only at the level of the message. It treats a rumour, trust shock or credibility breach the way infrastructure engineers treat a fault on a power grid: something that either dies out locally or runs away, depending on whether the network sits above or below a specific threshold.

Why does the power-grid analogy work for crisis communication? It works because cascading failure follows the same shape across very different systems — power grids, financial contagion, epidemics and information networks all share a threshold at which a local event either fizzles out or spreads. The analogy breaks down at the level of physics, since electricity is conserved as it flows while information replicates, but the structural shape of failure transfers cleanly.

How is this different from standard crisis communication planning? Standard planning optimises the message: the wording, the holding statement, the tone, and the timing. Cascade containment treats the network itself — its speed, its trust reserves, and its brittle points — as the thing being managed, on the basis that a well-crafted message still fails if it lands in a network that's already primed to run away.

What's the main limitation of this framework? It borrows the grid's protection philosophy and its method for proving value but deliberately doesn't borrow the grid's flow mathematics because information multiplies as it spreads while electrical current doesn't. It's also not a forecasting tool — it tells you how much margin you're operating with, not which shock arrives next or when.

How does this relate to trust-building work practitioners already do? Trust functions in this framework as carrying capacity – the property that determines how much a message can move across a relationship before that channel ignores, blocks or disbelieves it. That reserve has to be built before a crisis through credibility, fairness and legitimacy; it can't be generated by a holding statement once the crisis has already started.

What should a practitioner actually do with this before their next crisis? Start by checking which of the six containment protections you genuinely have versus which you've only assumed — most organisations find they have message discipline and little else. Then run a rehearsal designed to break something, since dormant failures like a stale contact list or a single-person credential only surface under real load, and finding them costs far less before a crisis than during one.

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

Dear {{ first_name | reader }},

I've always been drawn to the collaborative and knowledge management side of wikis, going back to my years as PR Manager at IBM, where we used them extensively to keep institutional knowledge from disappearing every time someone moved roles.

Over the years since, my own working material has scattered across folders, decks, and half-finished documents in a way that started to genuinely bother me. At some point I thought, I need to organise this properly, and I need some AI support to help me do it.

What follows is the result.

Introducing the Crisis Intelligence Wiki

This week I'm opening something I've been building hard over the past week: wiki.riskcomms.com, a public, free reference library for risk, crisis, and emergency communication.

It already holds dozens of documented case studies, the frameworks I use in client work, and the recurring patterns that connect them, and I'll keep adding to it for as long as the field needs it.

Crisis Communication Has No Shared Reference Layer

Our field has no shared reference layer. Every consultancy keeps its own case files. Every practitioner rebuilds the same mental models from scratch. When something like the GATE Model or a structured case study needs citing, there's rarely a stable public page to point to — just a slide deck buried in someone's drive or a memory of a conference talk.

Years of Scattered Notes, One Week to Structure Them

I started keeping my own notes on cases and frameworks years ago, the way most of us do — scattered files, half-finished decks, and things I'd half-remember and rebuild for every new client. What changed a week ago wasn't the material; it was deciding to actually structure it.

Once I started, it moved fast: a proper taxonomy of case studies, frameworks, themes, and organisational entities, each one cross-linked to the others, built and published within days rather than left as another someday project.

A Wiki Only Works if More Than One Person Writes It

A wiki only earns its name if more than one person writes it. I built the scaffolding – the structure, the templates, the first wave of entries – but a reference library authored entirely by one practitioner is really just my own notebook with a public URL attached to it. Whether this becomes useful infrastructure for the field depends on whether other people start treating it as theirs too.

What's Actually on the Wiki Right Now

The wiki currently holds 22 documented case studies spanning infrastructure failures, industrial disasters, aviation crises, and natural disaster response, plus three original frameworks (GATE, the Communication Nexus Framework, etc), a cross-cutting theme synthesising patterns across the case library, and a published audit tool co-developed with Ripple Research.

Every entry carries proper source attribution; where a figure can't be independently verified, the page says so rather than presenting it as settled fact.

Homepage of the Crisis Intelligence Wiki

Why I'm Keeping This Curated, Not Open-Edit

Public, collaborative reference resources have a well-documented failure mode: they either calcify (nobody updates them, and they go stale) or they degrade (anyone can edit, and quality erodes).

I'm deliberately avoiding both by keeping authorship curated rather than open-edit, at least for now, which means the wiki grows slower than a fully open one would but lets me hold every page to the same research standard before it goes live. "Curated" doesn't mean flawless, though, and I'd rather be corrected than pretend otherwise.

Case study view in the wiki

Where This Actually Saves You Time

If you're a crisis, risk, or emergency communicator, the wiki earns its place when you need a citable case study mid-engagement rather than a half-remembered anecdote or when you want to point a client or a junior colleague at a framework explanation that doesn't require you to write one from scratch every time. It's reference infrastructure for exactly the kind of work this newsletter exists to support.

How the Wiki Is Structured, and How It Gets Built

The wiki is organised around four content types and two publication types: case studies (documented incidents with timelines and lessons), frameworks (models like GATE and CNF), themes (cross-cutting patterns observed across multiple cases), entities (organisations and actors), audit tools (scored assessment instruments), and white papers (full-length analysis).

Everything cross-links through the others — a case study points to the frameworks applied within it, and a theme points to the cases that demonstrate it. The structure is the product as much as any individual page is.

Graph view navigation in the wiki

The workflow behind it is straightforward, even if it took some work to set up. I feed in raw material — source documents, notes, and sometimes just a rough outline of what I want documented.

AI organises that into structured markdown files following a fixed set of templates, one per content type, so every case study or framework follows the same shape and carries the same metadata.

Those files land in my Obsidian vault, where I review, edit, and link them before anything goes live. From there, the vault pushes to a GitHub repository, and Cloudflare Pages picks up the change and publishes it automatically.

Would you like to contribute?

The wiki grows if practitioners send me what they know. If you have a documented case study worth preserving, a framework you've built and want cited properly, or a correction to something already on the site, email me directly — phil@riskcomms.com (or reply here).

I'll review, credit you by name if you want that, and add it to the library.

That's this week's edition. Thanks for reading, and thanks in advance to anyone who sends something my way.

Philippe

Worth Attending + A Gift For You

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

SPONSOR

Want to get the most out of ChatGPT?

ChatGPT is a superpower if you know how to use it correctly.

Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.

Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.

Download the free guide

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

Dear {{ first_name | reader }},

Back in the late nineties, when I set up my first crisis room at IBM, we used flipcharts. One for decisions taken, one for actions in progress, one for what was done. Analogue, unglamorous, and completely effective. Everyone walking into that room — regardless of when they arrived or what they'd missed — could read the situation in thirty seconds without asking a single question.

I've pushed the same principle in every workshop since. You don't know who will be there when a crisis breaks. You don't know how long it will run. The person who starts the response is rarely the person who finishes it. Shift changes happen. People get pulled away. And the next person through the door needs to read the room, not get a briefing.

We moved to Kanban on a shared screen eventually. Different technology, same logic.
So when I read Kosuke Nakazawa's piece on what the City of Yokohama has built — 27 action cards on a wall, red to yellow to green — my first thought was: what a simple, practical solution. I sent him a DM saying so.

Then I sat with it a little longer and started feeling something closer to embarrassment. Because Yokohama had taken a principle I'd been improvising with for twenty-five years and turned it into a system. With institutional backing. And 27 actual cards.
Crisis communication still doesn't have that.

Here's what Kosuke documented — and what I've been trying to build in response.

This week's signal

I read a lot of emergency management literature. It keeps the practitioner side honest — and has the reliable side effect of reminding you how far crisis communication planning still lags behind crisis operations planning.

Kosuke Nakazawa's May 29 article on LinkedIn documented something the City of Yokohama has been doing quietly, without fuss: replacing its 100-page Emergency Operations Centre manual with 27 large A4 action cards, posted across an entire wall, tracked through a red-yellow-green colour system. Red: not started. Yellow: in progress. Green: done.

Anyone who arrived — experienced or not — could immediately see what needed doing and who was doing it. No briefings required. No waiting for the senior official who might be stuck in traffic on a Sunday morning.

And it made me ask an uncomfortable question: why hasn't crisis communication built the same thing?

Why it matters

Most organisations have crisis communication plans. Thick ones, often. Comprehensive, carefully written, regularly updated by someone who never has to use them under fire.

And most of those plans share the same fundamental flaw: they were written to be read, not to be used.

When a crisis breaks — really breaks — at 11 p.m. on a public holiday, when the comms director is on a plane and the deputy has the flu, what the first person in the room actually needs is not a plan. They need to know, in the next sixty seconds, what to do first.

The Yokohama model answers that question for EOC stand-up. What's the equivalent for crisis communication activation? For most organisations, the honest answer is 'nothing this good'.

That gap has consequences. Communication delays in the first hour are not neutral. They allow narratives to form without you. They signal institutional paralysis. They make everything that follows harder. The operational side of emergency management has understood this for decades. The communication side still hasn't fully caught up.

The background

The Incident Command System and its derivatives — ICS, NIMS, OSOCC, and the various national equivalents — have long embedded the principle of structured activation under uncertainty. The assumption is not that the right people will always be available. It is that the right system must work even when the wrong people are available.

This is not new thinking. It is hard-won doctrine from real disasters where the right people were not available, and the absence of structured activation cost lives.

Crisis communication has borrowed some of this language — activation protocols, spokesperson matrices, escalation tiers — but has largely failed to translate it into genuinely usable operational tools. Plans remain document-centric. Training remains scenario-centric. And when the moment comes, the person standing alone in the comms office is often left staring at a 47-page PDF on a password-locked SharePoint folder.

The Yokohama approach does something conceptually different. It doesn't trust the person. It trusts the system. The person just has to follow the wall.

Field notes

There's a benefit to systems like this that doesn't get discussed enough: they compress the experience gap.

I ran a crisis simulation recently with a mixed group of humanitarian professionals. Some had been through real incidents. They knew the procedures by heart and moved with the quiet confidence of people who had done this under actual pressure. Others were newer — capable, motivated, but having to find their way through each step, checking and second-guessing themselves while the scenario moved forward around them.

The gap between those two groups was visible. And in a real crisis, that gap has a cost.

A well-designed activation system doesn't eliminate experience. Nothing does. But it levels the starting conditions. The newer person isn't navigating from memory or instinct they haven't yet built. They're following the wall. The card tells them what to do, how long it should take, and what 'done' looks like. They spend their cognitive energy on the task, not on reconstructing the sequence from scratch.

The Yokohama model made exactly this design assumption — that the system had to work even when the most experienced people weren't in the room. That's not an edge case. In a large-scale disaster, it's close to a certainty. The experienced people are stuck in traffic, dealing with their own families, or already deployed somewhere else.
Build for the least experienced person likely to be first through the door. The veterans will adapt. The newcomers need the wall.

The core argument

The Yokohama model works because it solves the right problem.

The problem it solves is not knowledge. It's activation under stress by under-prepared people. The city officials who designed those 27 cards were not assuming their staff would be untrained. They were assuming, based on hard evidence from real drills, that even trained staff freeze when faced with an unstructured start state. The wall gives them the start state. The cards give them the sequence. The colour system gives them shared situational awareness without a single briefing.

Crisis communication has exactly the same problem. The person who picks up the phone when a journalist calls at midnight is rarely the crisis communications lead. It's whoever is on duty. The person who sends the first internal alert when the social media monitoring flags something at 6 a.m. is rarely the comms director. It's whoever is watching the dashboard.

These people need a wall. Not a manual. That's what I've been building.

What the evidence shows

Nakazawa's article documents the first blind drill Yokohama ran with the action card system. The results were instructive. The cards enabled smoother operations. But the drill also revealed something important: previous drills had always begun with the EOC already set up. Staff had practised responding to a crisis. They had never practised standing up the response infrastructure itself.

This is almost exactly parallel to what I find in crisis communication training. Organisations practise press conferences. They practise spokesperson interviews. Some practise dark-site activation. Very few practise the first fifteen minutes – the moment when a single generalist communicator has to decide whether this is real, who to tell, and what to say first.

The Yokohama approach isolates that specific gap and addresses it directly. The 27 cards don't cover the whole crisis response. They cover EOC stand-up — the critical window before the experts arrive.

Crisis communication needs the same scoping discipline. The first hour is its own problem. Solve that one first.

The complication

EOC activation and crisis communication activation are not the same thing, and a direct translation has limits worth naming.

EOC stand-up is largely procedural. Check the room. Contact the team. Verify the systems. These tasks are sequential, mostly independent of what the crisis actually is. You do them whether it's a flood, a fire, or a chemical spill.

Crisis communication is more contextual. The statement you draft for a product recall is not the statement for an industrial accident. The stakeholder map shifts. The tone shifts. The channels shift. Some of this cannot be pre-scripted without knowing the crisis type.

A card-based system for crisis communication has to work around this. Some cards are universal: decision gate, stakeholder notification, and holding statement. Others need conditional logic — if it's a reputational crisis, pull this card; if it's a safety incident, pull that one. And some things — the actual substance of what you say — will always require a thinking human in the loop.

The cards don't replace judgement. They create the conditions under which judgement is possible.

The practitioner angle

If you're reading this, you're probably not the generalist who needs the wall. You're the senior practitioner who builds the plan or trains the people who will use it.
Your plan may be good. The question is whether it's usable by someone who is not you, under conditions you cannot predict, in the first fifteen minutes of something going wrong.

Run this test. Pick the least crisis-experienced communicator in your organisation. Give them a scenario on a Saturday afternoon. Tell them a crisis is breaking. Don't brief them. Don't help them. Just ask them to activate the plan.

Watch what happens.

If they hesitate at the first step, the plan has a usability problem. If they pick up the manual and start reading from the beginning, the plan has a usability problem. If they reach for their phone to call someone more senior rather than starting any process, the plan has a usability problem.

The Yokohama drill found exactly this. Not because the staff were untrained, but because the activation mechanism itself had never been tested. The manual existed. The knowledge existed. The muscle memory for the first ten minutes did not.

AI rendering of draft crisis communication cards

The tool

After reading Nakazawa's piece, I went back to my own work and asked the same question he was asking about EOC stand-up: what does a crisis communicator actually need in the first ten minutes, and does anything we currently produce give it to them?
The answer was no. So I built something.

The Crisis Comms Activation Deck are 20 cards — plus one permanent banner card — that take a communicator from first alert to stand-down, regardless of their experience level. The logic borrows directly from Yokohama.

Red → Yellow → Green. Every card starts red. Flip it yellow when you start. Green when it's done. Progress visible at a glance.

Solo-operator first. Most crises begin with one person, alone, before reinforcements arrive. Every card has a Solo Mode section for exactly that situation.

Speed over completeness. The goal in the first ten minutes is not a perfect response. It is a “started” response. The cards are sequenced accordingly.

Decision gate logic. Card 01 asks one binary question: does this need an external response, or does it stay internal for now? The answer activates different card sets — preventing under-scaling and cutting off the paralysis that comes from not knowing where to begin.

Templates, not blank pages. Every card requiring a communication output carries a fill-in-the-blank template. Not because the template will be sent as written — it won't — but because a starting structure beats a blank page under pressure every time.
The specification is complete. The digital version is in development as a mobile-responsive web application, because crises don't wait for you to be at a desk.

Case in point

Yokohama's blind drill in January 2026 produced one finding worth sitting with. The official quoted in Nakazawa's article put it plainly: "People forget, so continuing regular training is essential. The cards are not a final product; they must be updated alongside the manual through repeated training and practical experience."

The cards are not a one-time fix. They are a living operational tool that degrades without maintenance.

The failure mode for crisis communication tools is almost always atrophy, not bad initial design. The plan that was excellent three years ago is now out of date because the social media manager changed, the CEO changed, the primary crisis monitoring platform changed, and nobody updated the card that says "call [Name] on [Number] and log into [Platform]".

Yokohama's commitment to updating the cards through repeated drills is not an afterthought. It is the operating model. That discipline is exactly what most crisis communication planning lacks.

Worth Attending + A Gift For You

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

Want to get the most out of ChatGPT?

ChatGPT is a superpower if you know how to use it correctly.

Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.

Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.

Download the free guide

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

Dear {{ first_name | reader }},

Over the last week I spent most of my time building something most consultants haven't thought about yet. Not a new service. Not a proposal.

A sovereign AI infrastructure — self-hosted, EU-based, and with no Big Tech dependency — so that when I use AI tools on client work, I can demonstrate exactly what runs where, what touches what, and what never leaves the secure system. Full transparency. Documented accountability. A transparency card for my own practice is in effect.

I'm a solo consultant. If I can build this, large organisations have no excuse for not having thought about it.

And then, while I was deep in server configurations and cron jobs 😵‍💫, a 91-page governance framework landed from CDAC Network, the Alan Turing Institute, and Humanitarian AI Advisory.

SAFE AI — the Standards and Assurance Framework for Ethical AI. Built for humanitarian organisations deploying AI in refugee camps and conflict zones.

Nobody in the private sector communications world seems to have noticed.

That's the story for this week.

Enjoy!

What is the SAFE AI Framework and why does it matter?

Somewhere in a refugee camp in northern Kenya, a humanitarian organisation is using an AI system to help determine who gets food assistance. The system makes recommendations. Staff act on them. And until last month, there was no shared sector standard for what happens when it gets it wrong.

That changed last May.

CDAC Network, The Alan Turing Institute, and Humanitarian AI Advisory published SAFE AI — the Standards and Assurance Framework for Ethical AI. Ninety-one pages. Two years in development. Fieldwork in Kakuma Refugee Camp with 193 community members across structured focus groups. Eight communities of practice. A roundtable at the UK Foreign Commonwealth and Development Office attended by ninety people from thirty institutions. Five authors spanning humanitarian leadership, AI accountability, responsible AI consultancy, and international policy.

The humanitarian sector built this because it had no choice. Its legitimacy depends on accountability to the people it serves. You do not deploy AI that shapes access to food assistance and then discover, when something goes wrong, that you have no governance architecture.

Read it. Then ask the question the framework forces you to ask. Does your organisation have the equivalent?

Does your organisation have an AI accountability gap?

The SAFE AI introduction states its problem plainly. Humanitarian organisations are deploying AI in contexts characterised by vulnerability, power asymmetry, and limited avenues for redress — without a sector-shared operational framework that translates global governance norms into appropriate safeguards.

Now swap the words.

Replace "humanitarian organisations" with "corporations". Replace "vulnerable populations" with "customers, employees, or the public." Does the sentence still hold?

It does. Almost perfectly.

The humanitarian sector named the problem publicly and built a framework to address it. Most private sector organisations have the same problem and have not named it at all. AI in production. People affected by those systems. No shared accountability architecture. The difference is not the stakes. The difference is the honesty.

Here is what SAFE AI actually operationalises: people whose lives are shaped by AI-influenced decisions have a right to know when automation is in play, a right to understand how it is shaping decisions about them, and a right to contest those decisions.

That is not a humanitarian principle. It is a communication principle. It applies every time your organisation's AI system touches a customer, a job applicant, a journalist, or a regulator.

What is an AI Transparency Card, and do you need one?

The central governance output of SAFE AI is something called a Transparency Card. A single, live document that records what the AI system does, the problem it was built to solve, who made the decision to deploy it, what risks were identified before launch, how those risks are monitored, who is accountable when something goes wrong, and how anyone affected can challenge the output.

It is not completed at launch and filed. It is updated continuously through the system's operational life. When the model is updated by the supplier, the Card is updated. When a new risk is identified, the Card is updated. When an incident occurs, the Card is updated. It is a living record, not a document you wrote once and hope nobody asks to see.

You do not have one. I am almost certain of this.

You have something adjacent. A responsible AI policy, legally drafted and comms-approved. A values statement. A vendor checklist from procurement, completed once, never revisited. The slide deck from the AI strategy offsite. None of that is a Transparency Card.

Think about the AI-related incidents of recent years. A hiring tool that discriminated. A customer service model producing harmful outputs. An automation system making consequential errors at scale with nobody able to explain how. In each case, the secondary damage — the narrative damage, the reputational damage, the regulatory exposure — came from the inability to say clearly: here is what the system did, here is who knew, here is the decision trail, and here is the governance.

The organisation that can produce that document on the day the story lands controls its own narrative. Not because the document exonerates them. Because it shows accountability was designed in, not improvised under pressure.

What does AI governance look like in practice?

I have run workshops on three continents this year. Different sectors, different sizes, different levels of AI maturity. And somewhere in every single session — whether I ask the question directly or it surfaces on its own — the same moment happens.

Someone raises the scenario of an AI system failing. Not catastrophically. Quietly. A wrong output acted upon at scale. And I ask: if that happens tomorrow, who do you call and what do you tell them?

The room goes quiet. Not because people haven't thought about it. Because the honest answer takes a moment to say out loud.

"We're working on it."

Every time. In organisations large and small, well-resourced and stretched, digitally mature and just getting started. The intention is genuine. The roadmaps exist. The conversations are happening at the right levels.

But good intentions and active roadmaps are not a named individual. They are not a documented decision trail. They are not a version history that tells a regulator or a journalist exactly what the system did, when it was last reviewed, and who signed off.

The gap is not a failure of awareness or ambition. It is a failure of infrastructure. And infrastructure, unlike intent, takes deliberate time to build — which is exactly why now is the right moment to start.

What does a complete AI Transparency Card cover?

The SAFE AI Transparency Card has six sections. None require humanitarian expertise. Each requires honest answers to questions your organisation should already be able to answer about every AI system in production.

Section 1 — System identity and purpose.

Name the system. Name the problem it was built to solve. Name every supplier involved and their contractual obligations. This sounds simple. But the framework flags a pattern it calls "adoption pressure" — the tendency to acquire AI systems because a supplier pitched them, a competitor announced one, or leadership wanted to do something with AI, and then construct the use case after the fact.

If the problem statement was written to justify a purchase rather than define a need, Section 1 will expose it. That exposure is useful. Better to surface it internally than have a regulator surface it for you.

Section 2 — Performance and known limitations.

Most private sector AI deployments measure accuracy. Few measure the cost of inaccuracy to the person on the receiving end. A customer service model with a 94% accuracy rate sounds adequate until you calculate how many customers per day are receiving wrong information, whether those customers are concentrated in a particular demographic, and whether anyone is tracking complaints back to AI-generated outputs.

Section 2 forces that calculation. It also requires documentation of what the system cannot do — the edge cases, the contexts where it degrades, and the failure modes. Suppliers will not volunteer this. You need to ask for it in writing.

Section 3 — Data sources and provenance.

The legal team will tell you the data use is covered by the terms of service. That is a legal answer. The communication question is different: if the customers whose data trained this model knew exactly how it was being used, would they feel the use was fair?

What is legally permissible and what people would consider reasonable are not the same question. The gap between them is where reputational exposure lives.

Section 4 — Risk monitoring and update triggers.

Your responsible AI policy written in 2023 does not account for a model your supplier quietly updated in March 2025. Your card needs two monitoring tracks. First, does your supplier contract require notification of material model updates? Second: has the environment around the model changed?

A customer communications tool trained before a product recall, a regulatory change, or a public controversy is operating in a different context than the one it was designed for. Someone needs to own the decision about whether the original risk assessment still holds. Name that person. Put it in the card.

Section 5 — Accountability and redress.

The framework demands a name. Not a team, not a function, not a shared inbox. The individual whose job it is to be called when something goes wrong. Alongside that name, your card needs a documented redress pathway. If a customer believes they have been harmed by an AI-influenced decision, what exactly can they do?

If your current answer is "contact customer services", that is not a redress pathway. A comms professional who cannot describe that pathway has no credible answer when it fails publicly.

Section 6 — Version history and decision log.

When a story breaks, the first question any journalist, regulator, or board member asks is: who knew what, when did they know it, and what did they do about it? A version log answers that question with evidence rather than reconstruction from memory.

Decisions that get logged are decisions that get made properly. When everyone knows the record will be reviewed, the review happens before the deployment, not after the failure.

What is the communications professional's role in AI governance?

Senior communications professionals are accountability infrastructure. Not metaphorically.

We are the function that explains what happened, to whom, on what basis, and what will be different. We are the function that gets called when the system fails and someone needs to say something true.

If the system fails tomorrow, what can you actually say? Not what you intend to say. What can you document, verify, and hand to a journalist at 9am?

Most senior comms professionals I talk to cannot answer that question cleanly. Not because they haven't thought about it. Because nobody has asked them to build the infrastructure that would let them answer it.

Building a Transparency Card is not a communications project. It is an organisational audit. But communications professionals should lead it — because they are the ones who have to answer for it when the system fails.

Start with the system that carries the most reputational risk. One card. Six sections. Honest answers. That is your starting position.

How does SAFE AI fit into global AI governance frameworks?

SAFE AI is part of a maturing global conversation about AI governance. The EU AI Act establishes risk-based obligations. The NIST AI Risk Management Framework offers structured assessment. The OECD Principles on AI emphasise transparency. UNESCO's Recommendation on the Ethics of AI provides globally endorsed norms.

What SAFE AI does that most of these do not is start from the people most affected — the communities and individuals whose lives are shaped by systems they often cannot see, question, or challenge. In humanitarian contexts, those are refugees and displaced persons. In commercial contexts, they are your customers, your employees, your job applicants.

The accountability logic is identical. The operating environment is different. The stakes are different. But the core question — can the people affected by this system understand how it works and challenge what it decides? — is the same.

The humanitarian sector answered that question with a framework. Your sector has had the question for longer and has mostly produced policies.

How do you know if your organisation's AI governance is adequate?

If your organisation's most consequential AI system failed tomorrow — not catastrophically, just quietly, at scale — what would you tell the first journalist who called?

Not the corporate line. Not the holding statement. What is the actual answer?

Where does the documentation sit? Who is the named individual accountable? What was the risk assessment? When was it last reviewed? Can you show the version history?

If you cannot answer those questions in the next sixty seconds, you do not yet have a governance architecture. You have intent.

Intent does not hold up in a crisis. Documents do.

How do you start building AI accountability in your organisation?

Inventory your AI systems.

List every AI system currently in production that affects customers, employees, or the public. Include AI features embedded in platforms you already use. Most organisations discover they have more than they thought. That inventory is Section 1 of your Transparency Card.

Name the person.

For each system on that list, name the individual — not the team, not the function — accountable when something goes wrong. If that name does not exist, that is the gap. Fill it before you need it.

Read SAFE AI.

The full framework is at cdacnetwork.org/safe-ai. It was written for humanitarian organisations in resource-constrained environments. Your organisation has more resources and fewer excuses. The Transparency Card template alone is worth the read.

Hope this helped!

Worth Attending + A Gift For You

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

Dear {{ first_name | reader }},

A quick word before we start.

No newsletter last Wednesday. I was in Belgrade, running a 2-day crisis simulation exercise for a humanitarian organisation client, and by the end of it I had precisely nothing left in the tank. You know how it goes, or if you don't, consider yourself lucky. 😅

Worth the wait, I think.

This edition covers something I've been wanting to write about for a few days: consultant colleague Rod Cartwright's 2026 Reputation, Risk and Resilience report.

Rod synthesises twelve major global risk and trust reports every year so the rest of us don't have to. This year's edition is, frankly, one of his best.

The findings are not comfortable. Neither are the questions they raise.

Let's get into it.

The 2026 Reputation, Risk and Resilience report

Every year, Rod Cartwright does something most of us lack the stamina to do: he reads everything. Twelve major global reports on trust, reputation, risk and resilience — Aon, AXA, the World Economic Forum, Edelman, Gallup, the Business Continuity Institute, and six more. Over 800 pages of data, distilled into one synthesis — the 2025 edition alone, was downloaded in 67 countries across six continents.

Rod and I go back. We share a professional obsession with the gap between what organisations say they are prepared for and what actually happens when things go wrong.

His annual report is one of the most honest mirrors the communications and risk profession holds up to itself. Not because it delivers surprises. Because it documents, with uncomfortable precision, the distance between awareness and action — year after year, across every sector and every geography.

As you read it, prepare to be challenged. Rod critiques what he calls the "rather supplicant refrain" of "why doesn't communication have a seat at the top table?" 

His hope is that this 2026 edition will help the communication function to "kick down the boardroom door and inhabit a seat that's been actively earned — through an attitude, mindset and substantive approach to practice that means it can't not be there".

The 2026 edition lands under the banner Assuming Agency and Achieving Balance in the Age of Angst and Rage. Rod is not backward in coming forward with his intentional choice of language. He describes his core term of agency "not as an illusory proxy for control, but rather as a central focus on intentional action and purposeful choice as the ultimate drivers of the real-world outcomes we aspire to achieve for ourselves and others".

Rod Cartwright's Assuming Agency framework – the conceptual spine of the 2026 report.

What follows are the ten themes Rod identifies, seen through the Wag The Dog lens – and what they actually mean if your job is crisis, risk, or emergency communication.

Have a cup or glass of your favourite drink and Enjoy!

The threat has gone economic. Your scenario planning is still thinking about tanks.

The World Economic Forum has placed geoeconomic confrontation at the top of its one-year risk rankings for the first time. Not armed conflict. Tariffs. Sanctions. Supply chain fracture. Trade weaponised as foreign policy.

This is the operating environment now. If your crisis protocols don't account for collapsing supply chains, evaporating market access, or the same message landing completely differently in Toronto than in Frankfurt — you have a structural gap. Geoeconomic scenario planning is not an add-on. It belongs at the centre.

Climate preparation is being quietly defunded by urgency. Someone should say something.

The WEF records something counterintuitive: absolute severity scores for all environmental risks fell over the two-year horizon. Yet extreme weather is still the #1 risk over ten years. Five of the top ten long-term risks remain environmental.

Economic and geopolitical urgency is crowding out climate work. The risks most likely to hit your organisation in the long term are receiving the least preparation investment right now. That is not a trend to monitor. It is a gap to name, escalate, and fix.

AI is now officially a risk, not just an opportunity. Your governance hasn't caught up.

Adverse outcomes of AI have moved from #30 to #5 in WEF's ten-year risk rankings — the single biggest cross-horizon jump of any risk in the report. Only 11% of risk experts believe public authorities are prepared for it.

Among senior communicators, Ipsos finds 57% use AI daily. Only 43% find it meaningful — down 15 points in a year. Just 11% believe their organisation's ethical policies are adequate.

Deepfake threats, synthetic media, AI-assisted disinformation — these are active crisis scenarios, not thought experiments. No response protocol for AI-sourced misinformation targeting your organisation? You are writing your crisis playbook for the wrong decade.

As Rod puts it:

Download, Print & Play Crisis Simulation Scenarios

When the ransomware hits, communications isn't the cleanup crew. It's the first responder.

Cyberattacks sit at #1 in the Aon survey for the third consecutive time. In the Asia-Pacific region alone, AI-driven deepfake attacks drove a 53% increase in social engineering incidents. Fraud claims in the same region rose 233%. In 2024, there were 2,593 ransomware attacks globally — up 15% on the year before.

Cyber incidents are narrative events. The reputational damage routinely exceeds the operational damage. Your crisis communication team needs to be inside the cyber response from the start — not summoned after the technical team has already made decisions that cannot be uncommunicated.

Workforce disengagement is a crisis indicator. Nobody is treating it like one.

Global employee engagement has fallen to 20% — its lowest level since COVID. Manager engagement has dropped nine points since 2022. Gallup puts the annual productivity cost at $10 trillion.

The BCI finds human-centred disruptions — safety incidents, health events, and workforce wellbeing failures — have topped the actual disruption index for the fifth consecutive year. They sit at 16th and 19th in forward planning. Aon records absenteeism as the #1 driver of actual financial loss. It ranks 39th in organisational risk registers.

Disengaged employees leak. They post. They step away from crisis protocols at precisely the moment those protocols depend on them. This is not an HR problem with communications side effects. It is a crisis preparedness problem that HR happens to be tracking.

To borrow a phrase from the report itself,

Strategic silence is a tactic. Organisations are treating it like a doctrine.

Eighty per cent of global corporate communications leaders expect ESG communications to become more tentative this year. In the US, 97%. The Ipsos Reputation Council calls it "strategic silence" — a deliberate decision not to engage on issues outside the organisation's core business.

Fine, for a period. The danger is when silence outlasts the hostility that prompted it. Walk into a crisis with no established position on the issues most material to your organisation, and you are not on neutral ground. You are on nothing.

You are no longer communicating to audiences. You are communicating to ever more insular tribes.

Edelman's 2026 Trust Barometer finds 70% of global respondents hesitant or unwilling to trust anyone who differs from them in values, information sources, or background. AXA: only 10% of the general population believe their country shares common values.

Richard Edelman, invoking John Donne's declaration that "no man is an island", warns that we are becoming "inflexible, intolerant, and incoherent in our cocoons".

Different communities will read the same message through incompatible frameworks and reach opposite conclusions. Crisis communication strategy needs audience segmentation built for fragmentation — not for the unified "general public" that does not exist and never did.

Your crisis response is built for one thing at a time. Nothing arrives one thing at a time.

International SOS: 57% of risk leaders say new risks are emerging faster than their organisations can manage. Forty-nine per cent report risk convergence has increased in the past twelve months. Aon: technology, trade, weather, and workforce pressures amplify each other. They do not queue politely.

International SOS pulls no punches:

A crisis response architecture built around single-event scenarios — one crisis, one team, one protocol — is structurally mismatched to the environment. Polycrisis response capability is not sophisticated. It is the floor.

Having a plan is not the same as being prepared. Most organisations have confused the two.

Only 14% of organisations quantify their exposure to their own top-ten risks. Eighty per cent of risk leaders know faster detection would improve outcomes. Twenty per cent can actually verify risk information at speed. The BCI finds organisations pouring preparation investment into risks that are not disrupting them, while the ones that are go unaddressed.

Can your organisation show — not claim, show — that its preparedness investment matches its actual disruption experience? If the answer is no, the plan is a document. Not a capability.

Every risk on this list eventually fails or succeeds at the human layer.

Rod saves what may be the most important theme until last — what he calls "the human risk blind spot and the human infrastructure deficit". The data are unambiguous.

Gallup: the failure of AI investment to deliver productivity gains is not a technology problem. It is a management problem. The BCI: cyber preparedness fails most often because of human error and disengaged staff, not technical inadequacy. The awareness-preparedness gap persists because of culture and judgement failures — not because risk registers are too short.

Organisations are returning to face-to-face relationship-building because every other form of connection is under strain. Human judgement, human relationships, and human resilience are the scarcest assets in crisis preparedness right now. More scarce than technology. More scarce than budget.

The quality of your crisis team — their training, their judgement, their wellbeing, and the trust they have built before anything goes wrong — is your most material preparedness variable. Everything else is infrastructure.

All figures in this article are drawn from Rod Cartwright's Reputation, Risk and Resilience: 2026 Edition, which synthesises twelve major global risk and trust reports published between mid-2025 and early 2026. It is free to download at rodcartwrightconsulting.com.

Read it. All of it.

Worth Attending + A Gift For You

trust,

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

SPONSOR

You Can’t Lead If You’re Stuck in the Middle

Most business owners don’t notice when it happens.

They start as the operator… and slowly become the center of everything.

The real risk isn’t being busy. It’s being required for everything.

BELAY’s new guide From Operator to Owner: How to Exit the Middle Without Losing Control shows you how to step out of day-to-day execution while keeping visibility, quality, and momentum intact.

BELAY matches you with U.S.-based Assistants who bring structure, follow-through, and operational clarity so work keeps moving without you at the center.

Download Operator to Owner for Free.

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

Dear {{ first_name | reader }},

In this week's edition of Wag The Dog, I'm focusing on three EU regulations that impact crisis communication, and the practitioners in Chicago, Sydney, and Dubai who assume none of them apply. (just kidding 😅)

On 7 May, the EU AI Act's Article 50 transparency deadline moved. Not by much. Enough to matter if you've been running a compliance timeline.

What follows is my attempt to map what the legislation actually requires, strip out the commentary that circulates on LinkedIn, and give you something actionable before December.

I've included the decision matrix. I've included the penalty tiers. And because I kept getting the same question from colleagues outside Europe, I've addressed extraterritorial reach directly.

My reading, not legal advice. Check the note at the bottom.

Philippe

Before we start… Ask Me Anything – 28 May

On Thursday 28 May at 5:00 PM GMT+2, I'm running the first Wag The Dog Ask Me Anything – one hour, no slides, no prepared remarks, 25 seats, no replay, Chatham House Rules from first question to last.

Bring the scenarios your organisation won't discuss out loud, the decisions with no good option, the gap between what the frameworks say and what the room demands at 11pm on a Sunday when the story is already moving, and I'll answer as honestly as I can. It's free.

The Regulatory Clock Is Ticking – And It's Set For December.

On 7 May 2026, the European Council and Parliament reached a provisional agreement that moved the EU AI Act's Article 50 transparency deadline from 2 August 2026 to 2 December 2026. Four months. If you've built a compliance plan around the August date, stop. Rebuild it around December.

That's the news. Here's what it means.

If You're Reading This From Chicago Or Sydney

Article 2(1)(c) of Regulation (EU) 2024/1689 is the mechanism: the Act applies to providers and deployers located outside the EU where the output is used or intended to be used within the Union. The GDPR model. Content travels; obligations travel with it.

A crisis communication team in New York generates a synthetic audio broadcast for a cybersecurity incident and distributes it to staff in Dublin and Madrid. The output landed in the EU. Article 50.4 applies. No EU address required.

A PR firm in Sydney drafts a press release for a global product recall distributed to European media. A senior director reviews and approves it before it goes out. The Act applies – the text reaches the EU – but the human review exemption kicks in. No label required. The exemption works across borders exactly as it does inside them.

A multinational in Tokyo deploys a customer service chatbot globally: EU users interact with it, Article 50.1 applies, and the disclosure must be there regardless of where the servers sit.

The practical answer for global teams: apply EU-compliant standards to all outputs, everywhere. Segmenting by intended geography is an operational minefield. Where distribution is digital, it's largely theoretical anyway.

What Article 50 Actually Requires

The AI Act runs to hundreds of pages. For communication professionals, four obligations inside Article 50 matter. Two fall on providers – the companies building AI tools, the OpenAIs and Anthropics of the world. Two land squarely on deployers. That's you.

Your organisation operates as a deployer. You're using AI in a professional capacity, not placing it on the market under your own trademark. The distinction shapes your legal exposure.

Chatbot Disclosure (Article 50.1) 

Any AI system designed to interact directly with people must disclose it's a machine. The provider builds the mechanism. You ensure it's active. In a crisis – under stress, time pressure, fear – your stakeholders are not at their most observant. Build the disclosure in. Don't rely on the exception.

Machine-Readable Marking (Article 50.2) 

Providers must embed detectable markers in AI-generated outputs. You don't build the watermarks. But if you generate content with a non-compliant tool, your organisation is exposed. Audit your toolstack now. Ask vendors directly about 50.2 compliance. If they can't answer clearly, that's your answer.

Deepfakes (Article 50.4, §1) 

If you generate or manipulate image, audio, or video that depicts real persons, objects, or events in a way that would falsely appear authentic, you must disclose it. The scope is broader than the word suggests: a synthetic CEO voiceover for an urgent briefing, a simulated crisis scenario for training. If it looks or sounds real and it isn't, label it.

AI-Generated Text on Public Interest Matters (Article 50.4, §2) 

If you use AI to generate text published to inform the public on matters of public interest, you must label it. Most commentary stops there. Read the next clause.

You don't, if two conditions are met: the text went through human review or editorial control, and someone holds editorial responsibility for publication.

If you use Claude, ChatGPT, or any generative model to draft a holding statement, a press release, or a risk advisory – and a human professional reviews it, edits it, and takes responsibility before it goes out – you are not required to label it AI-generated. Full stop.

The exemption was built for exactly this use case. It targets automated content that bypasses human judgment, not AI used as a drafting tool by professionals who own the output.

The corollary: automation without oversight loses the exemption entirely. A system that generates and publishes crisis updates to a public dashboard without human review must label those texts.

No human in the loop, no exemption. That's the line.

The Practical Matrix

The Penalty Picture

Article 99 of the AI Act tiers the penalties.

Violations of Article 50 transparency obligations sit in the lowest tier: up to €7.5 million, or 1.5% of global annual turnover, whichever is higher. Prohibited AI practices face €35 million or 7%. High-risk AI violations face €15 million or 3%.

€7.5 million or 1.5% of global turnover. Not trivial – and not the ceiling. That's the floor. Know the tier that applies to you.

The Other Two Clocks: NIS2 & DORA

NIS2 – in force since October 2024. In Member States that have transposed Directive (EU) 2022/2555, the obligations are already binding. For crisis communicators, the pressure point is Article 23: significant cyber incidents trigger a mandatory early warning within 24 hours of awareness, followed by a full notification within 72 hours.

That 24-hour clock starts at awareness. Not technical confirmation. Not legal sign-off. Awareness. If IT, legal, and communications don't have pre-aligned templates before an incident, your first 24 hours will be consumed by an internal argument. That's the failure mode I see most often.

NIS2 covers 18 sectors: energy, transport, banking, healthcare, and digital infrastructure in Annex I; manufacturing, food, chemicals, and postal services in Annex II. The sectors most likely to underestimate their exposure sit in Annex II. Check before assuming you're out of scope.

DORA – in force since January 2025. The Digital Operational Resilience Act entered application on 17 January 2025. The incident reporting sequence under DORA's Regulatory Technical Standards: initial notification within 4 hours of classification as a major incident, no later than 24 hours after detection; intermediate report within 72 hours; final report within one month. If your crisis templates say something different, fix them before you need them.

Between Now And December

Audit your tool-stack and document it. 

Assess every generative tool against Article 50.2 compliance. Ask vendors directly. If they can't answer, that's your answer. Document the responses – you'll need them.

Build the editorial documentation habit. 

The human review exemption only holds if you can show it actually happened. Log who reviewed, when, and who holds editorial sign-off. Those questions will arrive.

Build your labelling standards globally. 

Standardise the disclosure language – "This content was generated using artificial intelligence" – and apply it to all outputs regardless of geography. The EU AI Act's Code of Practice on marking will publish a common EU icon system. Final publication expected June 2026.

Integrate regulatory notification into your crisis plans. 

NIS2's 24-hour window and DORA's 4-hour initial notification are communication functions as much as technical ones. Pre-built templates and pre-agreed escalation paths must exist before you need them.

Three regulations. Three clocks. Two already running. The AI Act gives you until December – use the time. These regulations encode one principle: when AI produces content reaching the public, a professional is accountable for it.

That's not new. It's codified.

Worth Attending + A Gift For You

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

Worth reading

The Premium on Human Signals

In an environment where AI makes content incredibly cheap, we have to ask: what becomes more valuable? If everyone can generate a thousand words of "intelligence" in seconds, the market for generic information is effectively zero. Dan Lewis at The Comms Stack recently explored a specific edge in this new landscape: why human signals matter more in the AI content era.

Dan spoke with Cahill Camden, Founder and CEO of Novacast.ai, to tackle whether podcasts and actual conversations with actual people move the needle more now than ever before. The core takeaway is that as AI-generated text floods our feeds, human credibility becomes the ultimate premium. You cannot "operationalise" the credibility earned during a live conversation.

Why it’s important: If your communication strategy relies solely on volume and distribution, you are competing on a commodity. Real leverage now lives in the high-fidelity signals that AI can’t easily replicate like the nuance and authenticity of a podcast interview.

What’s next: Communicators should double down on platforms that prove there is a human behind the message. While we use AI to handle the speed/volume tasks, we must protect the "human signals" that preserve long-term institutional trust.

Also: If you’re not getting The Comms Stack in your inbox, it’s a brilliant resource for modern communicators, subscribe here.

SPONSOR

When it all clicks.

Why does business news feel like it's written for people who already get it?

Morning Brew changes that.

It's a free newsletter that breaks down what's going on in business, finance, and tech — clearly, quickly, and with enough personality to keep things interesting. The result? You don't just skim headlines. You actually understand what's going on.

Try it yourself and join over 4 million professionals reading daily.

Check it out

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

Dear {{ first_name | reader }},

On the evening of 15 March 2026, an unidentified luminous object appeared in the approach path to Cát Bi International Airport in Vietnam.

Two separate flight crews reported it. Four inbound flights diverted to Nội Bài. Two departures sat on the ground. Operations resumed after several hours of monitoring and verification, at 00:35 the following morning.

The object was never conclusively identified.

This is almost a perfect illustration of the communication problem that follows when you do exactly the right thing operationally. You detected an unknown. You acted. You protected the airspace. And then you had to explain all of that to a public that experienced the delays, saw nothing, and was told nothing conclusive.

That is the crisis communication challenge of the drone era. Not the object. The aftermath of appropriate caution, the necessary response to something you still cannot name.

I have been meaning to write this piece for some time. Cát Bi gave me the clearest example I have seen.

Enjoy!

Philippe

The incident

On the evening of 15 March 2026, two separate flight crews reported an unidentified luminous object in the approach path to Cát Bi International Airport in Hải Phòng, Vietnam.

One crew placed the object roughly 8 kilometres from the runway threshold at approximately 300 metres altitude. Air traffic control responded. Four inbound flights diverted to Nội Bài. Two outbound flights delayed. Operations resumed at 00:35 the following morning, after several hours of monitoring and verification.

The object was never conclusively identified.

Stay with that for a moment, because it contains almost everything you need to understand about drone crisis communication.

The operational response was not wrong. It was correct, proportionate, and legally required given what two independent flight crews reported: an unidentified object crossing the approach corridor of an active international airport. You act. You divert. You verify. You get an answer if the evidence supports one. In this case, it did not.

And now you have a different problem entirely.

The identification problem is the communication problem.

The planning gap

Most drone incident plans do not grapple with that. They prepare for a confirmed drone. They write holding statements about "suspected UAS activity." They map spokesperson roles and coordination chains.

They do not prepare for the operational requirement to act before identification is possible – and then explain that sequence to a public that experienced hours of disruption and received no conclusive answer about what caused it.

Cát Bi is not an edge case. It is a category.

Airports are the clearest example because the operational response is most visible: the diverted flights, the grounded departures, the departure board going blank. But the same logic applies to any facility that must decide, in real time, whether an unidentified object above its perimeter represents a nuisance, a crime, or a threat.

Prisons, power generation sites, water treatment plants, data centres in restricted corridors, sports venues during live events. The identification problem is identical. The communication failure that follows is structurally the same.

Beyond airports

What is your equivalent of the airspace call? The moment your facility takes protective action against an unidentified aerial object, the public clock starts. What goes out in the first fifteen minutes? Who signs it off?

If the object remains unidentified – or turns out to be something entirely benign – what does your explanation look like, and who has already agreed on the framing?

If you cannot answer those questions now, you will answer them under pressure, in public, with cameras pointed at your perimeter.

The same pattern runs from military-border spillovers in Romania and Estonia to civilian quad-copters near approach paths across Europe and Asia. The object changes. The communication problem does not.

The FAA receives more than 100 drone-sighting reports near airports every month. EASA has documented the same across European airspace – high volume, low confirmation rate, and a near-total absence of pre-agreed multi-agency communication protocols before incidents occur.

A meaningful share of those reports involve consumer quadcopters, balloons, birds at altitude, ultralight aircraft, and objects that are never definitively classified at all. The communicator's job in every single one of those cases is identical: explain what was seen, what was done, and what is still unknown – in that order, at that moment, without waiting for a conclusion that may not arrive.

That structure matters more than any individual message.

In practice, this is what breaks down. An airport spokesperson says "suspected drone activity." A law enforcement agency says "unconfirmed aerial object." The civil aviation authority says nothing for forty minutes. A local politician says "we are treating this as a serious security incident." A passenger films the empty gate and posts it with the caption "airport under attack."

By the time any official explanation reaches the public, the narrative has already run four laps without you.

Absent official answers, rumour fills the vacuum. EASA's post-incident analysis of European airport drone disruptions found that fragmented messaging between operators, authorities, and law enforcement consistently amplified public concern beyond what the incident warranted.

Drone incidents invite exactly this failure because the agencies are plural, the identification timeline is genuinely uncertain, and the visual material circulating on social media is usually dramatic and usually wrong.

What the first hour looks like

Your holding statement goes out when operational impact is confirmed – not when attribution is established. You do not wait for the object to be identified.

Four sentences: what was detected or reported, what operational action was taken, what has not yet been confirmed, when the public can expect an update. Ninety seconds to read. Months of pre-agreement to issue without internal argument.

The second task is preparing for the unresolved outcome. Not every incident ends with an answer. The Cát Bi object was never conclusively named.

Write the template for that scenario: the procedure responded to a credible report, the procedure produced a safe outcome, verification found no confirmed threat, investigation continues. That framing needs to exist before the incident, not be improvised when the answer fails to materialise.

The third task sits after the incident: a debrief across every agency that spoke publicly. Map the vocabulary drift, the attribution conflicts, the timing gaps. Those three things will repeat in the next incident if you leave them unexamined in this one.

The object determines the headline. The communication determines whether you still have authority when the next object appears.

If you found this article interesting, two resources belong in your kit.

The Drone Incident Communication Playbook gives you the operational communication structure for managing drone-related incidents — from first holding statement to post-incident debrief. The Silence Protocol puts your team inside a civil aviation crisis before one finds you.

Use code “DRONESUITE” at checkout on either or both. Twenty percent off. Valid for the next seven days.

Worth Attending + A Gift For You

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

SPONSOR

Write docs 4x faster. Without hating every second.

Nobody became a developer to write documentation. But the docs still need to get written — PRDs, README updates, architecture decisions, onboarding guides.

Wispr Flow lets you talk through it instead. Speak naturally about what the code does, how it works, and why you built it that way. Flow formats everything into clean, professional text you can paste into Notion, Confluence, or GitHub.

Used by engineering teams at OpenAI, Vercel, and Clay. 89% of messages sent with zero edits. Works system-wide on Mac, Windows, and iPhone.

Try Wispr Flow free

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

Dear {{ first_name | reader }},

I read something this week I cannot stop thinking about.

An article making the case that Jeff Bridges did not really act in The Big Lebowski. He just thought about his youth. The Dude was already in there. No performance required.

I love that film. And when I read that, I thought: there is more than one connection between Dudeism and crisis communication.

So this week, something lighter than usual: a man in a bathrobe, six lines of dialogue, and a doctrine that holds up better than most playbooks I have read this year. I think it still counts.

Enjoy!

Philippe

The Big Lebowski

The Big Lebowski is a film about a man who refuses to accelerate. Everything around him – the mistaken identity, the kidnapping, the nihilists, the rug – is pulling him toward a situation not of his making. He declines. He asks for the rug back, drinks a White Russian, and abides.

I started in crisis communication in 1994. I have spent decades watching organisations accelerate into crises they could have paced their way through. Most of them regret the speed more than the situation.

The classic failure is not bad messaging. It is good messaging delivered at the wrong pace, to the wrong audience, before anyone has asked the most important question in the room: does this moment actually require a response right now, or does the response become the story?

The Dude asks that question. He asks it differently.

Not every statement needs an answer

Not everything in a crisis is a fact. Not everything is an invitation to engage.

The noise arrives first – speculation, distortion, and stakeholder pressure from people who read a headline and drew a conclusion. The reflex is to correct all of it.

The Dude shrugs. He has assessed the statement and found it below the threshold. That is discipline dressed as indifference. Narratives you do not engage often die before the day is out. Narratives you correct are on record.

Protect one thing

In the middle of escalation, The Dude locates one thing and protects it.

In a crisis, that thing is the verified fact. Not the holding statement. Not the preferred executive framing. The fact. Everything else anchors to it.

Most teams scatter their attention across ten things and protect none of them. They surface with nine half-answers and a narrative someone else wrote.

The Dude holds the beverage.

Drop the performance

This is the line most crisis communications fail to mean.

Authenticity is not a value on a slide deck; it is a tactical position. Polished statements of concern are clocked immediately. Carefully hedged commitments to transparency are read for what they hedge.

The Dude does not perform. He is what he is, and there is nothing to map. No performance means no cracks.

Selective absence is an editorial decision

Not every call warrants a callback. Not every thread is a battlefield worth entering.

Crisis teams struggle with non-engagement because internal audiences read silence as inaction and inaction as failure. It is neither. Choosing not to respond is a decision about where the conversation happens, on what terms, and in what room.

The Dude opts out without apology. He does not justify the absence. He is not in.

Coherence is load-bearing

Messaging coherence is structural, not stylistic.

When the rug goes, the room goes with it. When messaging fragments – across spokespeople, channels, and timelines – credibility does not erode over days. It breaks the moment a journalist, or an employee, or a regulator, places two contradictory statements side by side.

The Dude returns to the same thing, again and again, because he knows what holds the room together. Not stubbornness. Narrative discipline. It is also the only thing standing between your organisation and a secondary crisis you built yourself.

Pacing is a decision

Speed without clarity produces retractions. Retractions produce a second crisis, one of your own manufacture.

Taking it easy is not passivity. It is the refusal to import other people's urgency. The Dude does not hurry. Urgency belongs to the people in the room around him; he does not pick it up. Crisis communicators who cannot separate their own pace from the noise around them issue corrections at 11pm.

The Dude does not issue corrections at 11pm.

Outlast, don't overpower

The goal is not to control the narrative. The goal is to still be standing when it burns out.

The communicator who abides – calm, selective, and consistent – does not overpower the crisis. They outlast it. Most crises burn out faster than the responses they generate.

Most corrections outlive the event that provoked them and become the permanent record. What you held, under pressure, is what remains.

Hold the beverage. 😉

Worth Attending + A Gift For You

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

Write docs 4x faster. Without hating every second.

Nobody became a developer to write documentation. But the docs still need to get written — PRDs, README updates, architecture decisions, onboarding guides.

Wispr Flow lets you talk through it instead. Speak naturally about what the code does, how it works, and why you built it that way. Flow formats everything into clean, professional text you can paste into Notion, Confluence, or GitHub.

Used by engineering teams at OpenAI, Vercel, and Clay. 89% of messages sent with zero edits. Works system-wide on Mac, Windows, and iPhone.

Try Wispr Flow free

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

Plastic toy figures. Rap videos. GTA-style missile edits. In the Iran–US–Israel confrontation of 2025 and 2026, some of the most strategically effective communication has looked nothing like communication at all. It has looked like content. Entertainment. A laugh.

For crisis, risk, and emergency communication professionals, this is not background noise. It is the new operating environment.

KEY TAKEAWAYS

• Wartime humour is not new — but its scale, speed, and reach are. States have used ridicule and comedy as propaganda tools for over a century. What changed is the delivery system.

• Your audience arrives pre-framed. By the time you issue a statement, many stakeholders already see the event through meme lenses that cast key actors as idiots, heroes, liars, or victims.

• Desensitisation is the slow, invisible risk. Repeated exposure to violence wrapped in humour gradually blunts emotional responsiveness — and makes it harder to signal genuine thresholds.

• Institutional humour around lethal events is almost always a bad idea. Know where your red lines are before the crisis, not during it.

• Build meme-aware monitoring into your intelligence picture. Recurring joke formats and caricatures shape trust and perceived competence just as powerfully as false statements.

Why Are Wartime Memes a Crisis Communication Problem?

Wartime memes are a crisis communication problem because they set the interpretive frame before any official message can land. By the time an organisation issues a statement, most stakeholders have already encountered two or three competing narratives — often packaged as entertainment.

Humour has always been a wartime tool. British broadcasters mocked Hitler on radio during World War Two. Cartoonists on both sides drew grotesque caricatures of enemy leaders. Comic books turned Axis powers into bumbling villains for young readers. The underlying psychology is not new. What changed is the form, the velocity, and the audience size.

In the current Iran–US–Israel conflict, this plays out across three distinct streams:

• Iranian propaganda uses Lego-style animations to portray Trump and Netanyahu as childish, incompetent figures pressing toy buttons — outmanoeuvred by cool, resourceful Iranian characters. Other clips mimic video games like GTA or Call of Duty, turning naval and missile engagements into something that looks like gameplay.

• US meme-style war content blends real strike footage with dramatic music and pop-culture references aimed at younger viewers — cinematic "hype" edits promoted by aligned media.

• Israeli trend-driven humour mixes nostalgic edits (1990s songs over military footage), "group chat" memes in which Iran's proxies "leave the chat" one by one, and snarky TikTok content designed to reassure domestic audiences while mocking adversaries.

The key insight for practitioners: the public's baseline understanding of a crisis is now heavily shaped by content that looks like entertainment, not information.

How Does Humour Work Psychologically During a Conflict?

Wartime humour works by giving people emotional distance from material that would otherwise be overwhelming — and that function is real, legitimate, and worth understanding.

Three mechanisms drive most of what we see.

Coping and anxiety reduction. Humour is a classic psychological defence. It lets people approach frightening material — war, escalation, mass casualties — without being consumed by it. Dark jokes and satirical edits offer a sense of symbolic control: we can laugh at this, so it cannot entirely dominate us. For citizens under fire, this can be genuinely protective. For those watching from afar, it keeps them emotionally engaged with a conflict that might otherwise feel too remote or too heavy.

Group bonding and identity signalling. Laughing together is powerful. Wartime memes encode in-group knowledge: shared references, slang, collective grievances. They answer the question, "whose side are you on?" without ever asking it directly. When Iranian Lego videos mock, or Israeli TikToks ridicule, they are not just attacking opponents. They are reinforcing an in-group identity — we are clever, resilient, and right; they are clumsy, corrupt, or ridiculous.

Ridicule as a weapon. Turning a feared actor into a clown lowers perceived threat and boosts confidence. It channels fear and anger into laughter. And crucially, it gives people an emotionally satisfying way to "win" against powerful adversaries without having to do anything at all.

Research on political memes published in the Journal of Social Media + Society (2023) found that exposure to politically charged memes influences attitudes and group identity formation — effects that were often underestimated by audiences who believed they were simply being entertained. (Source: Memeing Politics: Understanding Political Meme Creators, Audiences, and Consequences on Social Media — https://journals.sagepub.com/doi/10.1177/20563051231205588)

What Is the Hidden Long-Term Risk of Wartime Humour?

The most dangerous effect of wartime humour is not the misinformation it spreads — it is the desensitisation it quietly produces over time.

Repeated exposure to stylised or humorous depictions of violence reduces emotional responsiveness. This is not a theory. It is one of the better-documented phenomena in media psychology. When missile strikes, assassinations, and mass casualties are wrapped in Lego aesthetics, gaming metaphors, or snarky captions, audiences gradually:

• React with less shock or empathy when similar events occur.

• Begin treating extraordinary events as just another episode in a content stream.

• Reach first for memes and jokes — rather than information or considered response — when something new happens.

What was once unthinkable becomes ambient. War becomes content. Escalation becomes just more material to remix.

For risk communicators trying to signal genuine thresholds — this incident is different; this escalation creates new dangers — that is a serious structural problem. You are speaking into an environment where audiences have been trained, gradually and almost invisibly, to process war as entertainment.

As Philippe Borremans, founder of RiskComms and crisis communication specialist with 25 years of experience, observes: "You are not competing with misinformation. You are competing with a joke that landed three hours before your press release. The frame is already set."

How Should Crisis Communicators Respond to a Meme-Saturated Environment?

Crisis communicators operating in a meme-saturated environment need to make four practical shifts: develop memetic situational awareness, account for partial desensitisation in message design, establish clear internal humour policies before the next crisis, and resist the temptation to fight on the enemy's terrain.

Develop memetic situational awareness. Official briefings no longer sit at the top of the information hierarchy. They arrive late, into a conversation that has already formed. That means monitoring needs to expand beyond false claims. Track recurring joke formats. Map the caricatures circulating about your organisation, your partners, and at-risk groups. These shape trust and perceived competence just as powerfully as outright disinformation.

A 2018 study published in Science by researchers at MIT found that false news spreads significantly faster and wider than true news on social media — largely because novelty and emotional arousal drive sharing behaviour. Humorous content exploits both. (Source: The Spread of True and False News Online — https://www.science.org/doi/10.1126/science.aap9559)

Account for partial desensitisation. You will face overlapping audiences: people who are exhausted and use jokes to keep functioning; people who are sincerely engaged but whose first reaction to new information is an ironic one-liner; people whose empathy has been blunted by months of stylised violent content. Do not assume the facts will speak for themselves. They will not, in an environment where facts compete with spectacle. Mark thresholds deliberately. Use phrases like "for the first time…" and "this crosses a line because…". Keep bringing communication back to real people, real relationships, real consequences.

Set an internal humour policy before the next crisis. Agree in advance on when — if ever — your organisation will use humour in official crisis messaging. Which topics are strictly off-limits? How will you handle others' humour about you — ignore it, correct it, or reset with calm factual content? These calls cannot be made well at 2 a.m. on the first night of a major incident. Make them now, in writing, with sign-off from leadership.

Do not fight on the enemy's terrain. Weaponised humour will target institutions, spokespeople, and subject matter experts. The temptation is to clap back. That almost always helps opponents, not you. Focus on credibility and steadiness instead. You can acknowledge criticism without adopting its tone. Reserve whatever humour capital you have for lower-stakes, preparedness-oriented contexts where lightness serves genuine engagement — not for the middle of an active crisis.

A World Bank research paper on humour and risk communication (2021) noted that humour can be effective for increasing engagement and recall in preparedness contexts, but cautioned strongly against its use when the communication involves acute harm, vulnerable populations, or high-stakes decisions. (Source: Humor and Risk: Exploring a New Tool for Communication and Engagement — https://documents1.worldbank.org/curated/en/884411622710878616/pdf/Humor-and-Risk-Exploring-a-New-Tool-for-Communication-and-Engagement.pdf)

Should Crisis Communicators Ever Use Humour in Official Messaging?

Crisis communicators can use humour — but only in specific contexts, with clear institutional boundaries, and never around casualties, vulnerable groups, or high-impact events.

Grassroots humour by affected communities serves a real and legitimate psychological function. People under pressure need to cope. Mocking that impulse, or trying to co-opt it, is both paternalistic and professionally risky. The job of crisis communication is not to police how people feel. It is to provide clarity, credibility, empathy, and actionable guidance.

Institutional humour is a different matter. Official accounts joining the meme stream in the middle of a crisis create a distinct impression — and not usually the one intended. Humorous responses to serious harm can be rapidly reframed as callousness, arrogance, or incompetence, especially when clips are taken out of context and remixed.

Where humour does have a legitimate role: preparedness campaigns, low-stakes disruptions, and community engagement where the aim is to reduce anxiety and increase reach. Even there, the bar should be high. Know the audience. Test the message. And have a fallback.

How Do You Monitor for Meme-Based Narrative Risk?

Monitoring for meme-based narrative risk requires expanding your intelligence picture beyond traditional media tracking to include short-form video trends, recurring joke formats, and satirical caricatures circulating about your organisation or sector.

Practical starting points:

• Broaden your monitoring brief. Most media monitoring focuses on mentions, sentiment, and false claims. Add a specific category for meme formats and humorous framings. What joke templates are being applied to your issue? What caricatures of key actors are circulating?

• Watch micro-creators and war influencers. Creators and micro-influencers translate conflict into everyday language. They are often seen as more authentic than any institution. Know which creators your target audiences follow. Consider whether briefing or collaborating with trusted voices on preparedness and safety messaging is viable.

• Track emotional tone, not just accuracy. A piece of content can be factually neutral and emotionally devastating to your position. Recurring satire that portrays your organisation as slow, bureaucratic, or out of touch does real reputational damage — even if no individual claim is technically false.

• Integrate digital wellbeing awareness. Your audiences — and your own staff — may be saturated. Months of doom-scrolling, laughing, and dissociating take a toll. Pace your communications. Signpost resources. Acknowledge that oscillating between laughter and distress is a normal response to abnormal events.

What Does This Mean for the Crisis Communication Practice?

As of 2026, the emotional and cognitive baseline of crisis communication audiences is being shaped by content that blends war, comedy, and AI-generated spectacle at a scale no practitioner toolkit was designed for. That is the new ground. The question is what to do with it.

The lesson from Iran's Lego videos, US meme edits, and Israeli TikTok humour is not that crisis communicators should all become meme strategists. It is something more important and more difficult.

Our publics are not blank slates waiting for our next message. They arrive carrying emotional residue from content streams we did not create and cannot control. They have been trained — gradually, persistently — to process extraordinary events as entertainment.

Our job is to understand that baseline. To respect the coping function that humour legitimately serves for people under pressure. And to refuse to contribute to the desensitisation and trivialisation of real harm.

Communication that cuts through irony does not try to out-joke the jokers. It tells the truth, clearly and consistently, to people who are exhausted by spectacle and quietly ready for something real.

Frequently Asked Questions

What are wartime memes, and why do they matter for crisis communicators? Wartime memes are short-form digital content — videos, images, and jokes — that frame military or political events through humour, satire, or pop-culture references. They matter because they shape public interpretation of events before official communication arrives, often setting frames that are difficult to dislodge.

Are states actually using memes as deliberate propaganda tools? Yes. In the 2025–2026 Iran–US–Israel conflict, all three parties have produced or promoted content that blends humour, military imagery, and cultural references to shape domestic and international opinion. This continues a long tradition of wartime psychological operations, updated for short-form video platforms.

What is the difference between coping humour and institutional humour? Coping humour originates with affected communities and serves a real psychological function — it helps people manage fear and maintain group cohesion. Institutional humour is produced by organisations in official contexts. The two operate very differently. Institutional humour during acute crises involving casualties or vulnerable groups almost always creates more problems than it solves.

What is memetic situational awareness? Memetic situational awareness is the practice of actively monitoring the joke formats, satirical frames, and caricatures circulating around an issue or organisation — not just the factual claims and media coverage. It gives communicators a clearer picture of the emotional environment they are entering when they publish a message.

How does wartime humour contribute to desensitisation? Repeated exposure to violence wrapped in humour or gaming aesthetics gradually reduces emotional responsiveness. Audiences begin to experience extraordinary events as ordinary content. This makes it harder for communicators to signal genuine escalation or genuine danger when it arises.

Should organisations ever respond to memes that mock them? Rarely, and almost never in kind. Attempting to match the tone of satirical attacks usually amplifies them. A calm, factual reset is almost always more effective than a clap-back. If the meme contains a false factual claim, correct the claim. If it is purely tonal, consider ignoring it.

How do I build meme monitoring into my crisis intelligence function? Start by adding a short-form video and meme category to your existing media monitoring brief. Identify the platforms most relevant to your audience — TikTok, Instagram Reels, Reddit, Telegram. Track recurring formats and emotional framings, not just mentions. Treat humorous caricatures of your organisation with the same seriousness as factual attacks.

What role do AI-generated videos play in this landscape? AI-generated or AI-assisted content — including "AI slop" (low-quality synthetic media) — dramatically lowers the production cost of propaganda and satire. Anyone can now produce a passable Lego-style propaganda video in minutes. This means the volume of meme content in any future conflict will be significantly higher than what we have seen before.

How should I write crisis communications for a partially desensitised audience? Use explicit threshold markers ("this is different because…"), human-scale examples that re-anchor audiences to real consequences, and clear structure that makes it easy to extract the core message even for readers in a distracted or ironic frame of mind.

Where can I learn more about meme-aware crisis communication? The Wag the Dog Newsletter by Philippe Borremans covers emerging practice in crisis, risk, and emergency communication — including the intersection of AI, information warfare, and narrative intelligence. Visit wagthedog.io for current issues and archives.

References

• Memeing Politics: Understanding Political Meme Creators, Audiences, and Consequences on Social Media (2023) — https://journals.sagepub.com/doi/10.1177/20563051231205588

• The Spread of True and False News Online — MIT / Science (2018) — https://www.science.org/doi/10.1126/science.aap9559

• Humor and Risk: Exploring a New Tool for Communication and Engagement — World Bank (2021) — https://documents1.worldbank.org/curated/en/884411622710878616/pdf/Humor-and-Risk-Exploring-a-New-Tool-for-Communication-and-Engagement.pdf

• Do Memes Affect Our Political Ideas? — Universitat Oberta de Catalunya (2024) — https://www.uoc.edu/en/news/2024/memes-affect-political-ideas

• Synthetic Propaganda — Tracing an AI-Driven Political Operation (Podcast) — Bellingcat Stage Talk — https://rss.com/podcasts/bellingcatstagetalk/2744020/

Philippe Borremans is a crisis, risk, and emergency communication specialist with 25 years of experience and founder of RiskComms. He publishes the Wag the Dog Newsletter at wagthedog.io, covering emerging practice at the intersection of AI, information warfare, and crisis communication.

Dear {{ first_name | reader }},

I have sat in enough boardrooms to know the look.

You ask a senior executive about crisis preparedness. They nod. They gesture toward the binder, or the shared drive, or the consultant's report from two years ago.

"We have a plan," they say.

And they mean it as a closing statement – not an opening one. The plan is real. The readiness is not.

In this edition of the Wag The Dog newsletter, we’re going to look at the crisis readiness gap of the C-level suite.

Enjoy!

What the numbers actually say

The research on C-suite crisis leadership is not ambiguous. It is simply uncomfortable.

Only 28% of board directors understand their organisation's crisis plan "very well."¹ Just over a quarter – 26.5% – work in organisations that test their plans annually.² And roughly 70% lack a functioning cross-functional crisis team.³

This is not a picture of struggling organisations. It is the general state of crisis preparedness across sectors and geographies.

The consequences are quantifiable. Organisations that mishandle a crisis lose, on average, 10% of their stock price in the first week. Twelve months later, they are still sitting 15% below their pre-crisis baseline. Organisations where leadership performs well in crisis limit those losses to around 5%.⁴

That gap – 5% versus 15% – is not explained by the quality of the crisis plan. It is explained by the quality of the crisis leader.

What the plan cannot do

A crisis plan is a decision support tool. It captures what you know before things go wrong: protocols, contacts, escalation paths, and pre-approved language.

What it cannot do is make decisions in real time. It cannot read a room, or regulate emotion under pressure, or communicate credibly when the facts are still unclear and stakeholders are frightened. It cannot exercise judgment in a situation it has never encountered.

That is the leader's job.

Research identifies several capabilities that distinguish effective crisis leaders from ineffective ones. Standard executive competencies – the skills that drive performance in normal conditions – do not reliably transfer.

Crisis leadership requires something more specific: the ability to make sound decisions under conditions of high ambiguity and high stakes, to manage emotional responses (their own and others'), and to communicate in ways that build rather than erode trust.

Emotional intelligence, specifically, is a stronger predictor of crisis leadership effectiveness than most organisations acknowledge. Self-regulation under stress scores at β=0.485 as a predictor of performance. Empathy at β=0.361. Perceiving emotion accurately accounts for 18% of the variance in leadership effectiveness during a crisis.⁵

These capabilities do not appear in most crisis plans. And they do not develop through reading them.

The rehearsal problem

The research is direct on this: simulations, tabletop exercises, and scenario-based rehearsals are the primary mechanism for building genuine crisis capability in leaders.

They improve decision-making. They surface overconfidence – a documented contributor to poor crisis outcomes – before it becomes consequential.

They allow leaders to experience the conditions of a crisis (time pressure, information gaps, competing demands, emotional intensity) in an environment where the cost of failure is learning rather than loss.

The recommended minimum is annual. Many organisations do not reach it. Those that do often send the wrong people: middle management, communication teams, and operational leads. The C-suite is briefed on outcomes rather than put through the experience.

This is backwards. The highest-stakes decisions in any crisis will be made at the top. The preparation should match that reality.

Two tips to close the gap

For C-level executives:

Your crisis plan is not your crisis capability. It is a tool for your teams to use while you are doing the harder work of leading.

That work – making rapid decisions with incomplete information, holding the trust of stakeholders, and communicating clearly when clarity is scarce – requires practice under realistic conditions.

The next time your organisation runs a crisis exercise, be in the room. Not as an observer. As a participant. What you learn about your own responses under pressure will be more useful than anything written in the plan.

For communication professionals who advise C-suite executives:

You already know the gap exists. The harder challenge is getting permission – and budget – to address it.

Here is the framing that tends to land: this is not a communication training request. It is a risk management argument. A poorly led crisis costs organisations 10% of their stock value in week one, and the damage lingers for a year. A well-led crisis limits that to 5%.⁴

Frame it that way, and it stops being an HR conversation and starts being a board-level one.

The plan exists because someone understood that crises are predictable in shape, if not in detail. The investment in leadership capability needs to come from the same understanding: that when a crisis hits, no document will lead your organisation through it. A prepared leader will. An unprepared one will not.

Until next week,

Philippe

P.S. I am building a six-week cohort programme — Before the Crisis — designed specifically for C-suite executives who want to build this capability before they need it. Not for communications teams. For the people who will be making the decisions. If that sounds relevant, reply to this email and I will share more details when they are ready.

P.P.S. If you are a communication professional reading this, the programme is designed for your C-suite, not for you. If you have been trying to make the case for better executive preparedness, this might be the thing to forward.

Worth Attending + A Gift For You

I have a gift for you. Free tickets to the Natural Disasters Expo USA October 14–15, George R. Brown Convention Center, Houston – exclusively for Wag the Dog readers. This isn't a generic trade show. It's where the people responsible for keeping communities safe come to work through the hardest problems: climate change, ageing infrastructure, extreme weather events, and population growth. Government, emergency leaders, and private-sector innovators are in one room, focused on building resilience before the next disaster hits. Use this link. No cost, no catch – just your seat at the table.

Sponsor

Every headline satisfies an opinion. Except ours.

Remember when the news was about what happened, not how to feel about it? 1440's Daily Digest is bringing that back. Every morning, they sift through 100+ sources to deliver a concise, unbiased briefing — no pundits, no paywalls, no politics. Just the facts, all in five minutes. For free.

Read the newsletter trusted by 4.5 million fact-seekers.

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

UPDATE — 8 April 2026, 12:00 UTC+1: Since this article was first published, a ceasefire has been announced. It is a reminder of how quickly the situation on the ground can shift. The communication lessons, however, remain entirely relevant.

Key Takeaways

• Concealment is not a communication strategy. The Stargate facility's location was withheld from commercial maps — until an adversary made the reveal. At that point, there was no pre-positioned narrative, no established stakeholder relationships, and no communication posture to fall back on.

• Multi-partner infrastructure requires pre-assigned communication authority. When six named organisations share a crisis and none holds a mandate to speak for all of them, silence becomes the default. Silence is itself a message.

• Risk registers are not built for the speed of escalatory redefinition. A region's historical stability does not reliably predict future risk. State-actor conflict can reframe commercial infrastructure as a legitimate target within weeks. Most risk registers update annually.

• The hardest crisis to manage is one whose nature you cannot yet determine. A threat can be military, psychological, or both simultaneously. Crisis plans built around confirmed, categorised threats break down under ambiguity. Your framework needs to function before the threat resolves.

• A polycrisis has no single owner — and that is precisely the problem. Corporate exposure, investor anxiety, operational continuity, regional credibility, and information environment: five cascades, five stakeholder sets, five different communication clocks. If your crisis architecture assumes one point of ownership, it is not designed for the scenarios most likely to find you.

What Happened in Abu Dhabi on 3 April 2026?

On 3 April 2026, a video appeared showing satellite imagery of an apparently empty stretch of desert near the Abu Dhabi coastline — the kind of coordinates that appear on Google Maps as precisely nothing. It then switched to night-vision footage of the same location. The full footprint of a $30 billion data centre appeared. Visible. Mapped. Named.

The message overlaid on screen: "Nothing stays hidden from our sight, though hidden by Google."

That is not a military statement. It is a communication act — produced, distributed, and timed to land with specific audiences simultaneously. Iran had revealed the existence and precise location of the Stargate UAE facility, and in doing so, opened five simultaneous crises at once.

Sources: Tom's Hardware | TechCrunch

Why Does the Stargate UAE Facility Matter So Much?

The Stargate UAE project is not simply a data centre. It is a node in the global AI supply chain, and its scale makes a disruption almost unthinkable in practical terms — which is precisely why the threat lands so hard.

The facility broke ground in Abu Dhabi in March 2026 as a $30 billion initiative spanning ten square miles, with up to five gigawatts of planned power capacity. It is the largest dedicated AI infrastructure project outside the United States.

Built by G42 and operated by OpenAI and Oracle, with Cisco handling security and networking, SoftBank as financial partner, and NVIDIA supplying the GPU infrastructure, it sits beneath OpenAI's API — supporting training and inference workloads for GPT-based models and autonomous agent services.

OpenAI has described the facility as capable of serving up to half the world's population within a 2,000-mile radius. That is not a marketing claim. It is a statement of exposure.

A sustained disruption would not be a regional problem. It would cascade through every product and workflow running on that infrastructure, from enterprise API customers to consumer-facing services used by hundreds of millions of people.

What Is a Polycrisis, and Why Does This Scenario Qualify?

A polycrisis is not simply several crises happening at the same time. It is a situation in which multiple crises interact, amplify one another, and resist resolution by any single actor. The whole is genuinely more dangerous than the sum of its parts.

The Stargate threat qualifies because the video did not trigger one crisis — it set off a cascade across at least five interconnected domains, each with different stakeholders, different communication clocks, and different definitions of what a successful response would even look like.

What Were the Five Crisis Cascades Triggered by the Stargate Threat?

The Corporate Exposure Cascade

Six organisations are directly named in the Stargate partnership. Each has a different communication posture, a different stakeholder base, and a different relationship with the threat. None of them can resolve their communication challenge by acting alone. Their crises are structurally entangled.

Questions arose simultaneously about infrastructure viability, operational continuity, and the security of deployed assets. The consortium structure, which is a strength in normal operations, became a liability the moment silence became the default response.

The Investor and Financial Cascade

SoftBank alone has committed $100 billion to the broader Stargate initiative. Sovereign wealth funds and institutional investors sit behind the debt and equity structure of multiple Stargate sites globally.

When an asset is named in a threat video, the question reaching every investment committee is immediate: what is the current risk posture, and who is communicating on our behalf? Silence in those first hours is itself a message — one that destroys confidence faster than any explicit statement can rebuild it.

The Supply Chain and Operational Cascade

Data centres of this scale take years to plan, require specialised construction, and depend on supply chains already stretched by global semiconductor demand. They cannot be replaced quickly. There are no spare units sitting in a warehouse somewhere.

Every enterprise customer with GPU reservations, every developer with API commitments, every government agency with sovereign AI contracts tied to this facility had a business continuity question with no answer.

They were waiting for communication that, in many cases, was not coming — because no organisation in the consortium held a mandate to speak for all of them.

The Regional Credibility Cascade

The Stargate UAE facility was designed and announced as proof of concept: evidence that frontier AI infrastructure can be deployed in partner nations outside the US under sovereign governance frameworks.

It is the first deployment under OpenAI's "OpenAI for Countries" initiative. A credible threat against it is not just an operational risk. It challenges an entire strategic narrative.

Every government in conversation about hosting a similar facility is watching. Every minister who has committed political capital to an AI sovereignty partnership is quietly recalculating.

The Information Environment Cascade

One of the hardest challenges in this scenario is that the nature of the threat remains ambiguous. Is it genuine military intent? Strategic communication designed to create pressure and uncertainty? Something in between?

For crisis communicators, this is the most difficult position of all: managing the consequences of a threat whose character you cannot yet determine. Your response must function under multiple simultaneous possibilities. Most crisis playbooks are not written for that level of uncertainty.

What Did the Risk Register Miss?

For decades, Gulf states had remained largely insulated from direct conflict. That stability shaped risk assessments across the industry. The reasoning was defensible, as far as it went.

What it did not account for was the speed of escalatory redefinition — the process by which conflict between state actors reframes commercial infrastructure as a legitimate area of concern. That reframing can happen within weeks. Risk registers typically update annually. The gap between those two timescales is exactly where this scenario lives.

A second problem compounds the first. When a facility's existence and precise location are revealed simultaneously — in a produced, globally distributed video — you start from zero. No pre-positioned narrative. No stakeholder relationships established around the asset. No credible communication posture from which to respond. Everything must be built in public, under pressure, in real time.

Concealment is not a communication strategy. It is a deferred communication crisis. The reveal will come eventually. The only question is whether you control the timing — or someone else does. (Source: Fortune, March 2026)

What Are the Three Questions Every Organisation Should Be Asking Right Now?

These are not frameworks. Not models. Three questions that belong on the table in any organisation with physical or cloud infrastructure in a conflict-adjacent region — or any region where the risk environment can shift faster than the planning cycle.

Who speaks for the consortium?

In a multi-partner infrastructure arrangement, communication authority is rarely pre-assigned for crisis scenarios. Everyone waits. Silence fills the gap, and silence during a crisis is never neutral — it is always interpreted.

Establishing spokesperson mandate across multi-party arrangements is a governance question, not a legal one. Does your organisation have a clear answer? Have your partners agreed to it?

What does your stakeholder map look like at 3am?

Not the map built for business operations. The one that accounts for employees at a named facility when its coordinates have just been broadcast. Enterprise customers whose technical leadership is already on the phone to their cloud provider. Investors watching the news. Government partners trying to understand their own exposure.

Those audiences do not wait for office hours. Does your communication team know exactly who to reach, and how, within the first sixty minutes?

Where does your infrastructure actually sit — and does your communication posture reflect that?

Every organisation that has made location trade-offs — energy costs, connectivity, regulatory environment — without building a corresponding communication posture carries an unacknowledged liability. Not whether those trade-offs were right. Whether you have thought through what you would say if they became front-page news tomorrow.

Frequently Asked Questions About Polycrisis Communication

What is a polycrisis in the context of corporate communication? 

A polycrisis occurs when multiple distinct crises emerge simultaneously, interact with one another, and cannot be resolved by any single actor working alone. In the Stargate UAE scenario, corporate exposure, investor anxiety, operational continuity, regional credibility, and information warfare all escalated at the same time, each affecting the others.

What is the Stargate UAE project? 

Stargate UAE is a $30 billion AI infrastructure project that broke ground in Abu Dhabi in March 2026. Spanning ten square miles with up to five gigawatts of planned power capacity, it is the largest dedicated AI infrastructure development outside the United States. Partners include OpenAI, Oracle, G42, Cisco, SoftBank, and NVIDIA.

Why did Iran target the Stargate data centre? 

On 3 April 2026, Iran released a video containing satellite and night-vision imagery of the Stargate facility, accompanied by a message suggesting nothing remains hidden from their view. Analysts have noted the action appears to function simultaneously as a military signal and a strategic communication operation designed to create uncertainty and pressure.

What is escalatory redefinition, and why does it matter for risk registers? 

Escalatory redefinition is the process by which conflict between state actors reframes previously neutral assets — such as commercial data centres — as legitimate areas of concern or potential targets. It can happen within weeks. Because most risk registers update annually, there is a structural gap between how quickly risk changes and how quickly organisations adapt their assessments.

Why is multi-partner infrastructure particularly vulnerable in a crisis? 

When six organisations share a crisis and none holds a pre-assigned mandate to speak on behalf of all of them, the default response is silence. Silence is interpreted as evasion, incapacity, or indifference by different stakeholder groups. The governance structure that makes large infrastructure consortia efficient in normal operations becomes a communication liability in a crisis.

What is the difference between a security posture and a communication posture? 

A security posture addresses physical and cyber threats: what protections are in place, how they are monitored, and how incidents are contained. A communication posture addresses how an organisation responds to stakeholder questions: what it will say, who will say it, and when. An organisation can have excellent security and no communication posture — which is precisely what concealment strategies tend to produce.

How should crisis communicators handle threats whose nature is ambiguous? 

The key principle is that your response framework must function under uncertainty — not after it resolves. That means preparing holding statements for multiple threat scenarios, establishing clear decision thresholds for escalation, and identifying which audiences need which messages before you know the full picture. Waiting for certainty before communicating is itself a communication choice, and rarely a good one.

What is the "OpenAI for Countries" initiative? 

"OpenAI for Countries" is OpenAI's initiative to deploy frontier AI infrastructure in partner nations under sovereign governance frameworks. The Stargate UAE facility is the first deployment under this programme. A credible threat against it has implications not only for operational continuity but for every government currently in conversation about hosting a similar facility.

What should organisations do if their infrastructure is named in a threat scenario? 

Immediately activate your crisis communication team. Identify who holds spokesperson authority across all partner organisations. Assess which stakeholder groups — investors, customers, employees, government partners — need proactive communication within the first hour. Issue a holding statement that acknowledges awareness of the situation without confirming details that could create further exposure. Do not wait for the full picture before communicating.

How often should organisations audit their communication posture around sensitive infrastructure assets? 

As of 2026, annual audits are no longer sufficient in regions with elevated geopolitical risk. Quarterly reviews are a reasonable minimum, with immediate reassessment triggered by any significant shift in the regional risk environment. The audit should cover not only what the organisation would say but who is authorised to say it, through which channels, and to which audiences.

What is the broader lesson of the Stargate scenario for crisis communicators? 

Concealment is not a strategy — it is a deferred crisis. The organisations that navigate polycrises most effectively are not necessarily those that anticipated every scenario. They are the ones that had already asked which stakeholders needed communication, who held the mandate to deliver it, and what the organisation would say if its most sensitive assets became public. The Stargate scenario is a case study in what happens when those questions have not been answered before the crisis begins.

About the Author

Philippe Borremans is a crisis, risk, and emergency communication specialist with 25 years of experience advising organisations across the public and private sectors. He is the founder of RiskComms and author of the Wag the Dog Newsletter, a publication focused on crisis communication practice and the evolving risk landscape.

Dear {{ first_name | reader }},

This isn't a regular edition. No crisis debrief, no framework, no research findings. Just a tool and an honest explanation of why it exists.

A few weeks ago I built something. A structured AI skill, a set of instructions you install into Claude, that turns it into a sharp-elbowed press release editor.

The kind that refuses to start writing until it's asked you the uncomfortable questions. The kind that bans 'world-class' and 'innovative' and 'we are committed to transparency'. 🥱

The kind that forces you to answer who your real audience is – not the journalist but the person behind the journalist – and what exactly you want them to think, feel, or do differently.

It took a while to get right. It has been battle tested. I'm giving it away for free.

Here's why.

I wrote a few weeks ago about the line between intelligence and judgement, the argument that AI is absorbing the pattern-recognition, research, and formatting work of our profession, leaving the strategic calls to those who have the experience to make them.

Press releases were right there in my argument. I called it intelligence work. I'll go further: writing them is not a value-added activity. It never really was. It was always the thinking behind them that mattered; the audience, the angle, the judgement call about what to say and what to leave out.

But here's what I didn't say clearly enough: intelligence work done badly still causes damage.

A press release written by someone who doesn't understand the audience chain fails the organisation it was supposed to serve. A quote that sounds like it was written by a committee rather than a human being erodes credibility quietly, over time, in ways that are hard to trace back to the source.

The skill exists because there's a right way and a wrong way to do even the operational work. And because a generation of communicators (the one I wrote about, the one that deserves better tools than we had) shouldn't have to learn the hard way that 'innovative solutions provider' is a phrase that makes journalists delete emails.

So. Here it is.

How to install it (takes about a minute):

The skill is a .skill file designed for Claude (claude.ai). You'll need a free account if you don't have one.

1. Download the file from the link above.

2. Go to claude.ai and open Settings.

3. Find the Skills section and click Add Skill.

4. Upload the .skill file.

5. Done. Start a new conversation and ask Claude to help you write a press release. It will take it from there.

It works on the free plan. No technical knowledge needed.

One thing before you hit send on anything it produces.

This is a skill, not a ghostwriter. It will ask better questions than most briefing forms. It will flag the gobbledygook before it infects your draft. It will push you to think about the real audience, not just the journalist.

But the judgement is still yours. The answers to those questions are still yours. And whatever comes out the other end needs your eyes, your experience, and your sign-off before it goes anywhere near a journalist.

Generating and copy-pasting is not a communications strategy. It is a liability. The skill is there to make you think more clearly, not to think for you.

If you use it, I'd genuinely like to know what you think. Reply to this email. I read everything.

Have a great weekend.

Philippe

P.S. If you missed the long read on AI, judgement, and why I'm optimistic about the next generation of crisis communicators, it's here: https://www.wagthedog.io/p/long-read-what-ai-sacred-cows-and-the-next-generation-have-in-common — it's the context behind this tool.

KEY TAKEAWAYS

• Complacency is now a named risk. Control Risks explicitly identifies the "normalisation trap" — the structural dynamic where organisations mistake habituation to volatility for adequate preparedness.

• Compound crises are the new baseline. With 61 active state-based conflicts recorded in 2024 — the highest since World War II — planning for a single event at a time is the wrong question entirely.

• Your organisation is a target. Control Risks is explicit: activated societies direct their energy at specific companies, leaders, and institutions, not just governments.

• The governance gap creates communication paralysis. When regulatory frameworks are contested — AI, trade, technology — organisations without pre-built position frameworks hesitate. Hesitation costs as much as the wrong message.

• AI disinformation is already reaching your audiences. During the 2026 US/Israel/Iran conflict, AI-fabricated videos of missile strikes accumulated over 100 million views on X before effective intervention. Your stakeholders are already being targeted.

What Is the "Normalisation Trap" and Why Does It Matter for Crisis Communicators?

The normalisation trap is the risk that organisations become so accustomed to operating in volatile conditions that they mistake continued survival for genuine preparedness.

Control Risks names it directly in RiskMap 2026. It is not complacency in the loose, everyday sense. It is a structural dynamic: institutions calibrate their risk tolerance to the volatility they have already survived, then stop recalibrating. The environment keeps shifting. The calibration does not.

Organisations that navigated COVID, supply chain fracture, and geopolitical realignment often drew the wrong lesson from the experience. The lesson they took: we handled it. The lesson they should have taken: the conditions that produced that volatility are still intensifying — and the next wave will not look like the last one.

For crisis communicators specifically, the trap closes quietly. There is no alarm. The gap between the actual risk environment and your organisation's communication capability widens without fanfare, until the moment it becomes suddenly and painfully visible.

Why Are Three Separate 2026 Risk Assessments Saying the Same Thing?

Three major assessments published in early 2026 converge on a single underlying finding: governance is fracturing faster than organisations are adapting their communication capabilities.

The Global Catastrophic Risks 2026 report (Global Challenges Foundation, 2026), the Annual Threat Assessment 2026 (Office of the Director of National Intelligence, March 2026), and RiskMap 2026 (Control Risks, 2026) each approach the risk landscape from different angles. They arrive at the same place.

The numbers tell the story. Seven of the nine planetary boundaries — the scientifically defined thresholds within which human activity can safely operate across climate, biodiversity, freshwater, and other Earth systems — have now been breached (Global Challenges Foundation, 2026; PBScience, 2025). The ODNI recorded 61 active state-based conflicts in 2024, the highest number since World War II (ODNI, 2026). And Control Risks argues that the rules-based international order is not merely eroding — it is being deliberately replaced, with sovereignty now asserted through national security tools rather than multilateral frameworks (Control Risks, 2026).

Three different organisations. Three different methodologies. One common signal: the operating environment is significantly more unstable than most crisis plans reflect.

What Does "Cascading Crisis" Actually Mean for Your Communication Plan?

A cascading crisis is one where a single triggering event sets off secondary and tertiary failures across connected systems — and where the communication burden multiplies faster than response capacity can scale.

Most crisis communication plans are built around a separable-event logic. One thing goes wrong, you respond to that thing, you recover. Clean. Sequential. That logic is structurally inadequate for the current environment.

The GCF report describes environmental tipping points as cascading — one triggering another across climate, ocean, and biosphere systems (GCF, 2026). The same dynamic applies at the organisational level. A supply chain disruption coinciding with a cyber incident coinciding with an executive reputational crisis creates a communication challenge that no sequential plan was designed to handle. The compound interaction is where plans fail. Not the individual events — the collision between them.

The right planning question is no longer "what is the one thing most likely to go wrong?" It is "what happens when three things go wrong at once — and they interact?"

How Is Geopolitical Realignment Undermining Standard Crisis Messaging?

Messaging built on appeals to shared norms is losing reliability as a toolkit — because the norms themselves are actively contested.

Control Risks makes a stark argument in RiskMap 2026: the rules-based international order is being deliberately replaced, not simply challenged. Countries are imposing sovereignty through national security tools. Multilateral processes are giving way to bilateral and coalitional arrangements. Interests, not shared values, are the new architecture of international relations.

The consequences for communicators are immediate and operational. Appeals to international standards, institutional legitimacy, or established regulatory frameworks carry diminishing credibility when those frameworks are being renegotiated in real time. A message that lands well in one regulatory context may trigger hostility in another — increasingly because the contexts are diverging, not converging.

Control Risks pushes this further. As geopolitical alignment becomes consequential, your organisation's supply chains, operational territories, and partner networks carry associations that shape how your communications land across different audiences (Control Risks, 2026). Managing reputational risk and managing geopolitical positioning are no longer separable activities. Treating them separately now carries a higher cost than it ever has.

How Is AI-Generated Disinformation Changing the Crisis Environment?

AI-enabled disinformation is not a future risk for crisis communicators. It is the current operating condition.

The 2025 India-Pakistan conflict featured organised AI-generated content campaigns operating at speed and scale (ODNI, 2026). The 2026 US/Israel/Iran conflict went further. Iranian actors used AI tools — including Google's Veo 3 — to fabricate videos depicting missile strikes on Tel Aviv, downed US aircraft, and destroyed Israeli cities. Amplified across pro-Iran networks on X, these reached over 100 million views before effective intervention (BBC Verify, 2026; CEDMO Hub, 2026). AI-altered satellite images falsely depicting damage to US bases in Qatar were published by Iran's Tehran Times and circulated widely before fact-checkers could intervene.

That is the external threat. Most crisis plans at least acknowledge it exists.

The internal dimension receives far less attention. As AI tools are integrated into crisis communication workflows — drafting statements, monitoring sentiment, flagging emerging narratives — the question of accountability for AI-assisted decisions becomes genuinely complex. When an AI-informed decision proves wrong, existing governance frameworks rarely define who is responsible or how that error is attributed. Most organisations have not answered that question yet. They need to — before the next crisis, not during it.

What Three Actions Should Crisis Communicators Take Right Now?

All three 2026 assessments point to the same strategic response: build capability before the crisis that will reveal whether you have it.

Run a compound crisis exercise. If your most recent tabletop used a single-event scenario, you tested the wrong environment. The minimum useful exercise in 2026 combines a cyber incident, a reputational crisis, and a geopolitical disruption running simultaneously. The compound interaction — not the individual events — is where plans break down.

Map your activated society exposure. Control Risks identifies activated societies as targeting specific companies, leaders, and institutions — energised by perceived complicity as much as direct harm (Control Risks, 2026). This is not a social media audit. It is a threat intelligence exercise against your organisation's perceived stakeholder positioning. Who believes your organisation is complicit in something? Where are those audiences organising? What would trigger mobilisation?

Build pre-authorised frameworks for governance-gap scenarios. When regulatory and normative frameworks are unclear or contested — AI governance, supply chain integrity, geopolitical positioning — communication hesitation is as costly as the wrong message. Organisations with pre-developed position frameworks communicate faster and more coherently under pressure than those drafting positions in real time.

As Philippe Borremans, crisis, risk, and emergency communication specialist and founder of RiskComms, puts it: "The normalisation trap closes when organisations mistake their ability to cope with the past for preparedness for the future. The audit question is simple: are your current capabilities matched to the actual operating environment — or the one your last crisis plan was written for?"

FAQ

What is the normalisation trap in crisis communication? The normalisation trap, named by Control Risks in RiskMap 2026, is the structural dynamic where organisations calibrate their risk tolerance to volatility they have already survived — then mistake that calibration for genuine preparedness. It is dangerous precisely because it feels like stability.

Why is single-event crisis planning no longer adequate? With 61 active state-based conflicts in 2024 — the highest since World War II — and with cyber, reputational, and geopolitical pressures routinely intersecting, a single isolated crisis is increasingly the exception. Plans need stress-testing against compound scenarios where two or more events hit simultaneously.

What is a compound crisis tabletop exercise? A compound crisis tabletop runs two or more simultaneous crisis scenarios — for example, a cyber incident alongside a reputational crisis and a geopolitical disruption. It tests how your communication function performs when response resources are divided and narratives are competing for attention at the same time.

What does "activated society" mean in a crisis communication context? Control Risks uses the term to describe publics — consumers, employees, civil society groups — sufficiently energised around a cause or grievance to target specific organisations directly. Activation is triggered by perceived complicity, not just direct harm.

How is AI disinformation affecting crisis response in 2026? AI-generated content — including deepfake videos, altered satellite imagery, and algorithmically amplified narrative campaigns — is now a documented feature of conflict environments, as recorded in the ODNI's 2026 Annual Threat Assessment. It operates at a speed and scale that outpaces traditional fact-checking.

What is a governance-gap communication scenario? A governance-gap scenario is one where the relevant regulatory or normative framework is unclear, contested, or in active transition — AI regulation, trade policy, data sovereignty. Without pre-built position frameworks, organisations in these scenarios frequently hesitate, and hesitation itself becomes a reputational liability.

What should a pre-authorised crisis communication framework include? At minimum: defined topics requiring pre-approved positions, the process for activating those positions under pressure, clearly accountable decision-makers, and protocols governing AI-assisted response tools — including who can authorise their use and how errors in AI-assisted decisions are handled.

What are planetary boundaries and why should crisis communicators care? The Planetary Boundaries framework defines nine Earth system thresholds — including climate, biodiversity, and freshwater — within which human activity can operate safely. Seven of the nine have now been breached (GCF, 2026; PBScience, 2025). For crisis communicators, the relevance is compounding: environmental instability drives political conflict, displacement, infrastructure failure, and resource scarcity — all of which generate downstream communication crises.

How should organisations govern AI use in crisis communication decisions? Organisations should define, before a crisis occurs, which AI tools are authorised for use in response workflows, under what conditions, and who carries accountability when AI-assisted decisions prove incorrect. Leaving these questions unanswered until a live crisis is itself a governance failure.

How do I know if my crisis plan is built for the wrong environment? Ask when it was last stress-tested against a compound scenario. Ask whether it addresses AI-assisted response and AI-enabled disinformation. Ask whether it includes pre-authorised positions for governance-gap topics. If the honest answer to any of those is no, the plan needs updating — as of 2026, not eventually.

Sources:

• Global Catastrophic Risks 2026 — Global Challenges Foundation: https://globalchallenges.org/global-catastrophic-risks-2026

• Annual Threat Assessment 2026 — Office of the Director of National Intelligence: https://www.dni.gov/files/ODNI/documents/assessments/ATA-2026-Unclassified-Report.pdf

• RiskMap 2026 — Control Risks: https://www.controlrisks.com/riskmap

Philippe Borremans is a crisis, risk, and emergency communication specialist with 25 years of international experience. He is the founder of RiskComms and author of the Wag the Dog Newsletter, covering crisis communication strategy, AI's role in emergency response, and information warfare. Updated April 2026.

KEY TAKEAWAYS

• Most communication readiness tools measure confidence, not capability. In a polycrisis, that distinction is the difference between managing a situation and being consumed by it.

• The Polycrisis Communication Readiness Assessment evaluates organisations across five dimensions and 17 sub-dimensions, including crisis governance, narrative agility, and board-level influence.

• The methodology draws on the Design Science Research framework, the Becker et al. Eight-Phase Maturity Model, ISO 22361:2022, and the Sendai Framework.

• This is a self-assessment. The score you receive reflects the honesty you bring to it — not an external audit or certification.

• The most revealing results are rarely the high scores. They are the gaps you already suspected but hadn't named.

• The tool is free, the methodology is public, and the PDF report is downloadable.

Why Do Most Communication Readiness Tools Fail?

Most communication readiness tools are comfort exercises. The questions are vague enough that any experienced professional can work through them, tick "mostly yes," and walk away feeling prepared. That is not a small problem. It is the problem.

They measure self-perception, not operational reality.

In a polycrisis — where risks don't queue politely but pile in simultaneously, amplifying one another and collapsing the gaps between your response plans — the distance between thinking you are ready and being ready is enormous. It is the distance between managing the situation and being consumed by it.

Senior communication leaders need a tool that asks harder questions. Questions specific to the CCO function. Questions that map to how polycrises actually behave, not how textbook crises are supposed to behave.

"Communication readiness assessments that reward aspiration rather than demonstrated capability are not just useless in a polycrisis — they are actively dangerous. They create false confidence at exactly the moment organisations need honest intelligence." — Philippe Borremans, Crisis and Risk Communication Specialist, RiskComms, 2025

What Is the Polycrisis Communication Readiness Assessment?

The Polycrisis Communication Readiness Assessment is a structured self-evaluation tool for CCOs and senior communication executives, measuring capability across five strategic dimensions and 17 sub-dimensions. It produces a scored maturity profile, a gap analysis, strategic recommendations, and a downloadable PDF report.

One important clarification up front.

This is a self-assessment, not an external audit and not a validated research instrument.

What you get out depends entirely on what you put in. The instruction is simple: answer for your current state, not your aspirational one. That is harder than it sounds.

The methodology draws on several established frameworks:

• The Design Science Research framework

• The Becker et al. Eight-Phase Maturity Model

• ISO 22361:2022 (Crisis Management — Guidance)

• The Sendai Framework for Disaster Risk Reduction 2015–2030

The methodology page is publicly available for anyone who wants to examine how the scoring logic was constructed.

What Does the Assessment Actually Measure?

The assessment covers five strategic dimensions, each broken into sub-dimensions that reflect how polycrises behave in practice.

Strategic Competencies for Polycrisis Leadership

This dimension assesses geopolitical literacy, data and AI literacy, behavioural science acumen, digital diplomacy, and ethical reasoning. These are the analytical capabilities a CCO needs to read a complex threat environment accurately — before the first press inquiry arrives, before the first rumour circulates.

Organisational Resilience and Crisis Preparedness

Crisis governance, team capability, and organisational learning. Not whether your organisation has a crisis plan. Whether your infrastructure would hold under simultaneous, compounding pressure — the defining condition of a polycrisis.

Stakeholder Engagement and Trust Architecture

Stakeholder mapping, trust management, and multi-channel engagement. The question is not whether you have a stakeholder map. It is whether your engagement model is built for a world where institutional trust erodes in hours, not weeks.

Narrative Agility and Sense-Making

Rapid sense-making under ambiguity, narrative construction, and adaptive messaging in volatile contexts. In a polycrisis, the narrative environment is contested from the first hour. This dimension asks whether your team can move at the speed the situation demands.

Board-Level Influence and Strategic Positioning

C-suite and board access, communication as a strategic enabler, and resource mobilisation. If you cannot translate what you are seeing into language that moves decision-makers to act, the other four dimensions do not hold.

The assessment takes 30 to 45 minutes to complete.

What Do You Receive When You Complete It?

Completing the assessment produces a downloadable PDF report with four components.

• A maturity profile: Detailed scores across all five dimensions and 17 sub-dimensions, plotted on a five-level scale from Initial (Ad Hoc) through to Optimised (Resilient).

• A gap analysis: Where capability is weakest and where development investment will generate the most leverage.

• Strategic recommendations: Guidance tied directly to the gaps your results reveal, focused on what moving to the next maturity level actually requires.

• Progress tracking: Return, run the assessment again, and measure change over time. A single score is a data point. Repeated scores over months are evidence.

What Happened When the Tool's Author Ran It on Himself?

Transparency matters here. I ran the assessment on myself and scored 3.2 out of 5.0 — Level 3: Defined.

Level 3 means documented governance, standardised approaches, and established processes. Some sub-dimension scores landed where I expected. Digital Diplomacy and Cross-Cultural Agility, Stakeholder Mapping, Narrative Agility — twenty-five years of practice leaves marks.

The gaps were more useful than the scores.

Multi-Channel Engagement Strategy scored 1 out of 5. That reflects a deliberate choice: as a boutique consultancy, RiskComms does not run a dedicated digital engagement infrastructure. The score is accurate. But seeing it named plainly forced a question I had been avoiding: is that still the right choice?

Behavioural Science Acumen came in at 2 out of 5. Also accurate. I draw on behavioural science principles but have not systematised their application. The assessment's recommendation — to integrate these principles formally into message design, not simply reference them in strategy documents — is now on the list.

That is the real value of a diagnostic tool built this way. Not the high scores. The ones that make you uncomfortable.

Who Is This Assessment Built For?

This tool was built for CCOs and senior communication executives in organisations facing genuine complexity: government agencies, international NGOs, listed companies, utilities, health authorities, and financial institutions. Organisations where a polycrisis is not a thought experiment.

Communication consultants who want an honest picture of how their own practice maps against modern crisis demands will find it useful too. My own results are evidence of that.

One honest limitation: the maturity model assumes a certain operational scale. If you lead a team of three with limited crisis exposure, some sub-dimensions will not apply cleanly. The tool will not penalise you for it, but it is worth knowing before you begin.

How Does This Assessment Differ from Standard Crisis Audits?

Standard crisis audits are typically conducted by external parties against fixed criteria — process documentation, plan completeness, training records. They measure whether the infrastructure exists. The Polycrisis Communication Readiness Assessment asks a different question: whether the infrastructure would hold under the specific, compounding conditions of a polycrisis.

The distinction matters because polycrises expose the gaps between plans, not the gaps within them. Governance structures that function well in single-incident scenarios often fail when risks pile in simultaneously and amplify each other. A conventional audit would pass both scenarios. This assessment is designed to surface the difference.

How Should I Use My Results?

Use the gap analysis as your planning document, not the maturity score. The score orients you. The gaps tell you where to act.

Start with the sub-dimension scoring the lowest that sits within a dimension directly relevant to your organisation's threat environment. If your sector is facing active information warfare — financial services, healthcare, critical infrastructure — Narrative Agility gaps deserve immediate attention. If your challenge is getting communications onto the board agenda, Board-Level Influence and Strategic Positioning is where the leverage lies.

The progress tracking feature exists for a reason. Run the assessment now. Record your baseline. Return in six months with specific interventions behind you, and measure whether they moved the needle. That is what evidence-based professional development looks like.

Frequently Asked Questions

What is a polycrisis, and why does it require a different communication approach? A polycrisis occurs when multiple separate crises interact simultaneously, amplifying each other in ways that compound their individual impacts. The term was popularised by historian Adam Tooze to describe compounding global disruptions. Standard crisis communication frameworks are typically designed for sequential or single-event scenarios. A polycrisis demands different capabilities: faster sense-making, higher narrative agility, and governance structures that do not collapse under simultaneous pressure.

How long does the assessment take to complete? Between 30 and 45 minutes for most users. The time varies depending on how closely you examine sub-dimensions where you are uncertain about your organisation's current state.

Is this assessment appropriate for smaller communication teams? The maturity model assumes a certain operational scale. Smaller teams may find that some sub-dimensions do not apply directly to their context. The tool will not penalise you, but a team of three with limited crisis exposure should approach the results with that caveat in mind.

What frameworks does the assessment methodology draw on? The methodology draws on the Design Science Research framework, the Becker et al. Eight-Phase Maturity Model, ISO 22361:2022 (Crisis Management — Guidance), and the Sendai Framework for Disaster Risk Reduction 2015–2030. The full methodology is publicly available.

Can I share my results with leadership or my board? Yes. The downloadable PDF report is structured for that purpose. The gap analysis and strategic recommendations sections are particularly suited to board-level conversations about investment priorities.

How is this different from a standard crisis communication audit? A conventional audit measures whether crisis infrastructure exists — plans, processes, training records. This assessment measures whether that infrastructure would hold under the specific, compounding conditions of a polycrisis. The distinction matters because polycrises expose the gaps between plans, not within them.

What does a Level 3 maturity score mean in practice? Level 3 — Defined — means documented governance, standardised approaches, and established processes. It indicates a functioning foundation with identifiable gaps at the sub-dimension level. Most experienced communication functions operating without recent systematic development tend to score in the Level 2–3 range.

How often should I retake the assessment? Every six months, or following a significant crisis response, a major organisational change, or a substantial development investment. Repeated scores over time are evidence. A single score is orientation.

Is the assessment free? Yes. The tool is free, the methodology is public, and the report is downloadable. There is no paywall.

What should I do if my results raise questions I cannot act on alone? The gap analysis will identify where development investment has the most leverage. For questions about what your results mean for your specific context, or how to build the case for investment at board level, Philippe Borremans is available for consultation via RiskComms.

About the Author

Philippe Borremans is a crisis, risk, and emergency communication specialist with more than 25 years of experience advising governments, international organisations, and private sector clients. He is the founder of RiskComms and the author of the Wag the Dog Newsletter, which covers crisis communication strategy, information warfare, and the role of AI in emergency response. Philippe developed the Polycrisis Communication Readiness Assessment, the Universal Adaptive Crisis Communication (UACC) Framework, and the Minimum Viable Crisis Communication (MVCC) Framework. He is based in Europe and works internationally.

What Is Wrong With Most Mis- and Disinformation Triage Frameworks?

Most triage frameworks are built around the content rather than the audience — and that single structural error undermines every decision that follows.

Pick up almost any mis- and disinformation response guide published in the last five years. Attend a crisis communication workshop. Open a corporate crisis management playbook. At some point, someone will show you a triage grid. Three criteria — usually some version of Reach, Impact, and Virality. A colour-coded prioritisation matrix. Red for act now, green for watch and wait.

It will look clean. Logical. Reassuringly systematic.

And it will start in the wrong place.

Every standard framework treats the mis- or disinformation itself as the unit of analysis. What the narrative says. How far it has travelled. How dangerous it sounds in the abstract. That framing has always been a mistake.

Crisis communication starts with the audience and ends with the audience — not the message, not the channel, not the threat. Strip those audience questions out of your triage model and you don't have a triage model. You have a content catalogue.

The colour-coded grid tells you how big the fire is. It tells you nothing about whether your hose reaches the building — or whether the people inside would open the door if it did.

What Are the Six Dimensions That Actually Drive Response Decisions?

Effective mis- and disinformation triage requires six audience-centred dimensions — none of which can be assessed without a specific audience at the centre of the analysis.

The standard three-factor model isn't wrong because it chose the wrong three factors. It's wrong because it frames every factor as a property of the content rather than a property of the relationship between content and audience. The six dimensions that actually drive response decisions are these:

Harm Proximity. Not abstract harm potential — how close is this audience to an irreversible action, right now? The same dangerous content carries a completely different risk profile depending on whether it is reaching people three days before an event or thirty minutes into one. The audience's position in time is the variable. The content is constant.

Narrative Velocity. Not current virality — trajectory within the specific communities that matter to you. A story accelerating fast inside a small, high-influence network of community leaders is more operationally dangerous than one with ten times the reach in a general population that is not primed to act. Velocity only means something relative to an audience.

Source Authority Index. Who is carrying this narrative to this audience? An anonymous account is one thing. A retired health official, a community elder, a religious figure with deep roots in the affected population — these are force multipliers. The same mis- or disinformation travelling through different sources requires fundamentally different responses, because the audience's perception of source credibility is what determines believability.

Credibility Capital. What credibility has your organisation accumulated with this specific audience before this moment? Where institutional credibility is high, a correction lands. Where it is low — or where it has been actively eroded — a correction can amplify the problem. It can draw attention to the narrative, trigger reactance, or confirm the suspicion that authorities are hiding something. Credibility is earned through behaviour over time. You cannot manufacture it in the middle of a crisis.

Correction Reach Asymmetry. Can you actually reach the audience the mis- or disinformation is reaching — through channels they use, in a voice they find credible, faster than the narrative is spreading? If the answer is no, "correct quickly" is not a strategy. It is wishful thinking. The asymmetry is not about your platform reach in the aggregate. It is about whether you have a credible path to these people, right now.

Narrative Stickiness. Does this content confirm something this audience already believes or fears? Mis- and disinformation that slots into pre-existing beliefs does not need virality to be dangerous. It needs to land once. Then it sits inside the audience's existing narrative architecture and reinforces itself every time a related event occurs. Stickiness is not a property of the content. It is a property of the fit between content and audience worldview.

Why Does the "Monitor Only" Box Create the Most Risk?

The green "monitor only" box on most triage grids is the most dangerous cell on the matrix — because it creates the illusion of a decision without requiring one.

Every piece of mis- and disinformation that became a full crisis started there. Low reach. Seemingly low harm. Someone made a reasonable call to watch and wait. Then something shifted — a political event, a media pickup, one influential community voice sharing it with a simple comment — and what had been background murmur became the dominant narrative within hours.

Monitoring without a defined threshold for reassessment is not caution. It is observation without a plan. You are watching something grow without having decided what it needs to do before you act. And without tracking whether the audience's relationship to that content is changing — a shift that often happens well before the reach numbers move.

The fix is simple and enormously consequential: attach a trigger to every monitoring decision. Monitor only, unless X changes in the audience environment, at which point it re-enters triage automatically. That defined condition is the mechanism most existing frameworks are missing.

How Should a Triage Framework Adapt to Different Operational Contexts?

The same intellectual foundation should produce two very different tools — one for a solo practitioner under pressure, one for a team with time to think.

The solo public information officer sitting alone in an emergency operations centre at 2am cannot run a six-dimension analytical model while managing media calls and updating the incident command team simultaneously. That person needs something fast, binary, and defensible. Three questions. Two minutes. A clear output.

The communications team at a well-resourced organisation has thirty to sixty minutes before a major response decision. They can — and should — work through the fuller analysis. The richer the picture, the more precisely calibrated the response.

The quick version collapses the six dimensions into three fast audience-centred signals: Harm Proximity, Narrative Velocity, and Credibility Capital. These are the three where getting it wrong fastest causes the most damage. The output is not a colour. It is a named response posture — Intercept, Contain, Correct, Monitor, Stand By — each with one or two specific action prompts attached.

The team version unpacks all six dimensions, adds the deliberation that Correction Reach Asymmetry and Narrative Stickiness require, and produces a fuller picture: response posture, channel recommendations, rationale summary, and — critically — a reassessment trigger. A defined condition that sends the item back through triage automatically if circumstances change.

Few existing frameworks include that last element. A built-in clock. Because the assumption that a triage decision made at 9am is still valid at 3pm is how mis- and disinformation crises become communication disasters.

Why Did the Old Frameworks Work — and Why Do They Fall Short Now?

The original triage models gave non-specialists a vocabulary and a starting point. That was genuinely valuable. In a world where most communicators had no mis- and disinformation framework at all, a simplified grid was better than nothing.

That moment has passed.

The information environment is faster, more fragmented, and more adversarial than anything those frameworks were designed for. Audiences are more polarised. Institutional credibility is more fragile. And the consequences of a miscalibrated response — correcting the wrong thing, to the wrong people, through the wrong voice — can be as damaging as the mis- and disinformation itself.

As Philippe Borremans, crisis, risk, and emergency communication specialist and founder of RiskComms, has observed in his work with practitioners across sectors: "The upgrade isn't more criteria. It's a different foundation. It starts with the audience. It ends with the audience. Everything in between is method."

The information environment has moved on. The triage models need to follow.

Frequently Asked Questions

What is a mis- and disinformation triage framework? A triage framework is a decision-support tool that helps crisis communicators prioritise which false or misleading narratives require an active response and which can be monitored. Most existing frameworks assess the content itself — its reach, virality, and potential harm. Audience-centred frameworks assess the relationship between that content and a specific audience, which produces more operationally useful decisions.

Why is audience the right starting point for mis- and disinformation triage? Because the same piece of mis- or disinformation can be high-risk in one audience context and low-risk in another. A false claim reaching a community that already distrusts official sources, days before they must make an irreversible decision, is categorically different from the same claim reaching a general population with no immediate stakes. Content analysis alone cannot capture that difference.

What is Correction Reach Asymmetry and why does it matter? Correction Reach Asymmetry describes the gap between where mis- or disinformation is travelling and where your correction can realistically reach. If the false narrative is spreading through channels you cannot access — in a voice your target audience does not find credible — then issuing a correction may have no effect, or may actively draw attention to the false claim. Assessing this asymmetry before you act is essential.

What is Narrative Stickiness in crisis communication? Narrative Stickiness refers to how well a piece of mis- or disinformation fits an audience's pre-existing beliefs, fears, or worldview. Sticky narratives do not need wide reach to be dangerous — they need to land once with the right audience. Because they confirm what people already suspect, they are resistant to correction and tend to resurface repeatedly after related events.

What is a reassessment trigger and why should every triage decision have one? A reassessment trigger is a defined condition — a change in the audience environment, a media pickup, an influential voice sharing the narrative — that automatically sends an item back through triage for re-evaluation. Without one, a monitoring decision made at 9am remains in force at 3pm even if the situation has fundamentally changed. A built-in clock is the mechanism most existing frameworks are missing.

What is Credibility Capital in the context of mis- and disinformation response? Credibility Capital refers to the accumulated trust an organisation holds with a specific audience at the moment a false narrative emerges. Where credibility is high, corrections tend to land. Where it is low, corrections can backfire — amplifying the narrative, triggering reactance, or reinforcing suspicion that authorities are concealing something. Credibility cannot be built during a crisis. It must be earned before one arrives.

What is the Source Authority Index? The Source Authority Index assesses who is carrying a false narrative to a specific audience. An anonymous social media account and a respected community elder are not equivalent carriers of the same content. The latter is a force multiplier — and requires a fundamentally different response strategy, because the audience's perception of source credibility shapes how believable the narrative becomes.

How does the six-dimension model differ from a standard three-factor triage grid? Standard triage grids assess content properties: reach, virality, harm potential. The six-dimension model — Harm Proximity, Narrative Velocity, Source Authority Index, Credibility Capital, Correction Reach Asymmetry, and Narrative Stickiness — assesses the relationship between content and a specific audience. None of the six dimensions can be usefully evaluated without a defined audience at the centre of the analysis.

Is a detailed six-dimension framework practical under operational pressure? Not always — which is why the model produces two tools. The quick version collapses the six dimensions into three fast signals (Harm Proximity, Narrative Velocity, Credibility Capital) for use by a solo practitioner under time pressure. The full version is designed for teams with thirty to sixty minutes before a major decision, and produces a richer output including response posture, channel guidance, and a reassessment trigger.

What does a "named response posture" mean in practice? Rather than outputting a traffic light colour, an audience-centred triage tool outputs a named posture: Intercept, Contain, Correct, Monitor, or Stand By. Each posture comes with one or two specific action prompts attached — making the transition from analysis to action faster, clearer, and more defensible under operational conditions.

Philippe Borremans is a crisis, risk, and emergency communication specialist with 25 years of experience working with organisations across the public, private, and non-governmental sectors. He is the founder of RiskComms and the author of the Wag the Dog Newsletter, which covers emerging practice in crisis and emergency communication.

Dear {{ first_name | reader }},

Last year I was in Sharjah, delivering a crisis, risk, and communication masterclass with Divakar Pandian for the team at BEEAH Group — one of the UAE’s most ambitious sustainability and innovation organisations.

Good room. Sharp questions. And somewhere between the sessions, Divakar mentioned he was working on something bigger. That something is now published.

In February 2026, Resilience Next released GRS 1001-RN:2026 — the Global Resilience Standard. Version 1.0. Available for adoption. With a certification scheme attached.

I was one of the contributors, alongside nearly forty practitioners from across the industry. I am not a neutral observer, so take that into account.

But I would not be writing about it if I did not think it mattered for the people who read this newsletter.

Happy reading.

Why a new standard?

The world is not short of frameworks. ISO 22301 covers business continuity. ISO 31000 covers risk. ISO 27001 covers information security. There are national standards, sector-specific requirements, regulatory regimes — and enough acronyms to fill a whiteboard twice over.

The problem is not that we lack standards. The problem is that the existing ones were built in silos, and organisations consistently fail in the gaps between them. The cyber team does not talk to the communications team until something is already on fire.

The risk register and the crisis plan live in different departments, updated on different schedules by people who rarely meet. ESG gets bolted onto governance structures rather than woven into them. Nobody connects the social and environmental stakes of a disruption to the operational response before it happens.

GRS 1001-RN:2026 tries to close those gaps. It proposes a single Global Resilience Management System (a GRMS) that brings enterprise risk, business continuity, cybersecurity, crisis management, health and safety, crisis communications, ESG, AI governance, and third-party risk management into one integrated model. One governance structure. One audit. One certification.

Version 1.0 will have gaps. It is explicit about that — Version 1.1 is already scoped. But the architecture is sound, and for the first time in my experience, crisis communication is not an afterthought.

Communication Is Inside the Framework. Actually Inside It.

For most of the history of this profession, crisis communication has been a response function. Something happens, you call the communications team, and they draft a statement. The argument that communicators should be embedded in risk planning and governance (not just activated when things go wrong ) has been made many times. It has rarely been implemented.

GRS 1001-RN:2026 bakes it in from the start.

The Resilience Management Committee, the governing body overseeing the entire framework, explicitly seats a Communications lead alongside risk, continuity, cybersecurity, legal, and HSE functions.

The Respond pillar requires a documented Crisis Communications Plan as a certified deliverable, not an annex. And the standard applies the same four-stage lifecycle to communication that it applies to every other discipline: Define, Design, Embed, and Improve.

That last point deserves a moment. Communication in a crisis is not improvised. It is designed, practised, and continuously improved — or it fails. The standard treats it the same way it treats a recovery time objective or a supply chain redundancy plan.

If you have spent years making the case for a permanent seat in governance, this framework gives you a new argument. A certified one.

The Crisis Activation Threshold: Worth Stealing

One of the most practically useful concepts in the document is the Crisis Activation Threshold; pre-defined, documented criteria that determine when an Incident becomes an Emergency, and when an Emergency becomes a Crisis.

Most organisations do not actually have this. What they have is a vague understanding that someone will escalate to the CEO when things get bad enough. The result is predictable: slow decisions, the wrong people in the room, reactive communication, and the first hours spent arguing about whether this is actually a crisis rather than managing it.

The standard requires those thresholds to be documented and aligned with the organisation’s Risk Appetite, approved at board level before anything goes wrong.

For communication professionals, this changes the game. Pre-defined activation criteria mean pre-positioned messages, pre-approved spokespersons, and pre-tested channels. You are not starting from zero at the worst possible moment.

Why the Existence of This Standard Matters

Standards shape professions. ISO 22301 turned business continuity from a specialist niche into a certified discipline with auditable requirements and board-level visibility. ISO 27001 did the same for information security.

When a standard exists with a governance structure and a certification pathway, it changes what organisations ask for when they hire, what procurement teams require when they buy services, and what boards expect when they ask about resilience.

GRS 1001-RN:2026 includes crisis communication as a certified requirement. For those of us who have long argued that communication is the connective tissue of every crisis response — the mechanism by which organisations maintain legitimacy when everything else is breaking down — that is a meaningful shift.

This will not be the last time I write about this standard in Wag The Dog. As certification pathways open, capacity building develops, and early adopters begin implementation, I will keep you updated.

Where to start?

The standard is available for free download at globalresiliencestandard.org. Read it. Share it with the people in your organisation who own risk, continuity, and governance. Start mapping where your current frameworks align, and where the gaps are.

The conversation is just beginning.

SPONSOR

Attio is the AI CRM for modern teams.

Connect your email and calendar, and Attio instantly builds your CRM. Every contact, every company, every conversation, all organized in one place.

Then Ask Attio anything:

• Prep for meetings in seconds with full context from across your business

• Know what’s happening across your entire pipeline instantly

• Spot deals going sideways before they do

No more digging and no more data entry. Just answers.

Start your free trial →

Stay resilient, Philippe Borremans www.riskcomms.com

_{Transparency & Disclosures}

_{AI Transparency: In alignment with EU AI Act requirements, please note that AI technology was used in the research, drafting, and/or image generation for this edition. All strategic analysis, professional opinions, and final editorial oversight are conducted exclusively by the author. Affiliate Disclosure: Some links in this briefing may be affiliate links. I only recommend tools and services I use personally or have vetted for professional efficacy. Professional Advice: This newsletter is for educational and informational purposes only and does not constitute legal or professional crisis management advice. © 2026 RiskComms FZCO. All rights reserved.}

Key Takeaways

• The 2026 Middle East conflict revealed that financial instruments — specifically war risk insurance — can shut down global shipping more effectively than military force.

• A compromised prayer app with 5 million users delivered psychological operations messages that bypassed state internet blackouts entirely.

• False misinformation about nuclear testing spread within 17 minutes of Iranian seismic activity — well before any official body could respond.

• Reactive fact-checking fails once an audience is emotionally captured. Pre-bunking — warning people before manipulation arrives — is significantly more durable.

• Crisis intelligence today requires monitoring insurance markets, freight rates, and P&I club bulletins — not just social media.

• Single-channel communication strategies are a structural vulnerability. Redundancy is no longer optional.

What Does the BadeSaba Prayer App Reveal About Modern Channel Strategy?

The channels themselves are now contested infrastructure. US and Israeli operators compromised the BadeSaba prayer app — 5 million users — to deliver messages urging Iranian military personnel to lay down arms. No broadcast tower. No leaflet drop. The app bypassed state internet blackouts entirely.

If you are still thinking about channel strategy in terms of social, paid, owned, and earned media, that framework no longer covers the territory. Your audiences are operating inside a cognitive battle-space whether they have registered that fact or not. Any application with sufficient reach and user trust is, in principle, a potential delivery mechanism for someone else's message.

The lesson is not that you should abandon digital channels. It is that channel selection now requires a threat assessment, not just a reach-and-frequency calculation.

Why Does Misinformation Travel Faster Than Your Approval Process?

The 17-minute window does not care about your legal review cycle. During the February 2026 earthquakes in Iran, social media was flooded with claims of covert nuclear tests within 17 minutes of the seismic event — before the Comprehensive Nuclear-Test-Ban Treaty Organisation had issued anything, before a single professional seismologist had examined the data.

The claims were wrong. The correction came hours later. By then, large parts of the audience had moved on, carrying the wrong version.

Your crisis communications plan almost certainly assumes time to convene a response team, draft a statement, clear legal, and issue something considered. That assumption is now a liability.

The question is not whether you can respond in 17 minutes. Most organisations cannot, and trying to build a process that fast invites rushed errors. The real question is whether you have pre-approved holding positions and decision frameworks that allow a designated person to publish something — anything accurate and stabilising — without calling anyone first.

Pre-approved language is not a compromise. It is the difference between being present in a narrative and being absent from it.

How Did Insurance Markets Close the Strait of Hormuz?

The weapon was financial. The battlefield was actuarial. The Strait of Hormuz did not close primarily through military force during the 2026 escalation. It closed through insurance. The P&I clubs — Skuld, Gard, London P&I — issued 72-hour cancellation notices for war risk cover. Additional War Risk Premiums jumped from 0.15% to approximately 1% of hull value. Maersk and Hapag-Lloyd suspended operations or rerouted around the Cape of Good Hope. Daily vessel transit fell 86% below baseline.

The signal was present in the insurance markets before a single commercial vessel was touched.

For crisis communicators, this reframes what intelligence actually means. Monitoring social media for brand mentions is one thing. Monitoring Additional War Risk Premium surges, freight rate movements, and P&I club bulletins as early-warning indicators of escalation is another entirely — and that second capability is the gap currently costing organisations their response windows.

Organisations with supply chains, energy dependencies, or logistics exposure to contested regions need to integrate financial market signals into their crisis intelligence picture. This is not the job of the communications function alone. But the communications function needs to be in the room where those signals are being read.

Why Is Fact-Checking Losing the Information War?

Reactive debunking fails once an audience is emotionally captured by a narrative. The correction arrives too late. And it often reinforces the original claim through simple repetition — stating the false claim in order to deny it gives the false claim another impression. You are playing whack-a-mole against opponents who do not need to be accurate. They need to be first, and emotionally resonant.

The US Department of Defence has reached the same conclusion. It is now building information inoculation into warfighter training — warning personnel in advance that manipulation is coming and showing them what the techniques look like before they encounter them. The operational term for this is pre-bunking.

The DEPICT framework structures the threat landscape into six categories: Discrediting, Emotional language, Polarisation, Impersonation, Conspiracy promotion, and Trolling. Each category has recognisable signatures. Each can be taught.

The principle scales well beyond military applications. Any organisation whose workforce, clients, or stakeholders are likely to be targeted during a regional crisis — one affecting supply chains, energy prices, or financial markets — has a pre-bunking problem it has probably not yet addressed.

What Does Information Sovereignty Mean for Your Message Reach?

A message can be accurate, timely, and well-crafted — and still not arrive. Iran's internet connectivity dropped to 4% during the 2026 crisis. The state shifted from reactive censorship to a whitelist model: only authorised services permitted, with permanent digital trails left by anyone who strayed outside them.

This is the extreme end. But it sits on a spectrum that includes EU regulatory frameworks, platform content moderation decisions, and national emergency communications protocols that can be activated faster than most organisations anticipate.

Single-channel communication strategies are a structural vulnerability. Build redundancy now. Not after the first failure, when it is too late to establish alternative channels, verify alternative audiences, and test whether those channels actually work under pressure.

What Does the 2026 Middle East Crisis Tell Us About Crisis Communication Readiness?

The 2026 Middle East crisis is a convergence event — and most crisis communication plans were not built for convergence. Kinetic and narrative operations ran in parallel. Insurance markets functioned as strategic weapons. Seismic monitoring networks were exposed to coordinated manipulation. Trusted civilian applications were repurposed for psychological operations.

None of this fits neatly within a communications function designed for a slower, more linear world. If your crisis communication plan predates AI as an operational reality, predates information warfare as standard statecraft, or predates financial instruments as conflict tools, the gap between your plan's assumptions and current conditions is significant.

Peter Sandman, the risk communication researcher whose outrage model has shaped the field for decades, has long argued that perceived risk is as much about fairness, trust, and control as it is about probability and magnitude.

The 2026 events add a new dimension: the information environment itself is now a managed, contested space. What audiences perceive is being actively shaped by actors whose goals are not yours. That changes the communication calculus entirely.

Frequently Asked Questions

What is the 17-minute problem in crisis communication? The 17-minute problem refers to the documented speed at which false narratives — in this case, claims of covert nuclear testing during Iranian seismic activity in February 2026 — can saturate an audience before any official body has verified or responded to an event. Most organisational approval processes cannot match this speed, making pre-approved holding statements and delegated publishing authority essential tools.

How did war risk insurance become a strategic weapon in the 2026 Middle East conflict? When P&I clubs including Skuld, Gard, and London P&I issued 72-hour war risk cover cancellations, shipping operators faced uninsurable exposure. Additional War Risk Premiums spiked from 0.15% to approximately 1% of hull value. Major carriers suspended or rerouted operations, causing an 86% drop in daily vessel transit through the Strait of Hormuz — achieved without direct military action against commercial shipping.

What is pre-bunking and why is it more effective than fact-checking? Pre-bunking involves warning audiences in advance that manipulation attempts are coming and showing them what those techniques look like. Research and practice — including the US Department of Defence's warfighter training programmes — indicate that inoculating people against manipulation before they encounter it is significantly more durable than reactive corrections, which arrive after emotional capture has already occurred and can inadvertently amplify false claims through repetition.

What is the DEPICT framework? DEPICT is a framework for categorising information warfare techniques: Discrediting, Emotional language, Polarisation, Impersonation, Conspiracy promotion, and Trolling. It provides a structured vocabulary for recognising and teaching resistance to manipulation, applicable in both military and civilian organisational contexts.

Why is single-channel crisis communication a structural vulnerability? Filters — sovereign, regulatory, or platform-based — can sever communication between organisations and their audiences without warning. The 2026 crisis demonstrated that both state censorship and insurance-driven operational shutdowns can create information blackouts rapidly. Redundant channels, established and tested before a crisis, are the only reliable mitigation.

What should crisis communicators be monitoring that they probably are not? Beyond conventional media monitoring, crisis communicators with supply chain or energy dependencies should track Additional War Risk Premium movements, P&I club advisories, freight rate indices, and cargo route disruption signals. These financial and logistical indicators often move ahead of public narrative — they are the early warning layer that most communications functions currently lack.

How do compromised civilian applications change crisis communication planning? When applications with millions of legitimate users can be repurposed as psychological operations delivery platforms — as the BadeSaba prayer app was in 2026 — channel trust becomes a variable rather than a constant. Crisis communicators need to audit the digital infrastructure their audiences rely on and build resilience against the possibility that familiar channels are delivering adversarial content alongside legitimate information.

What does information warfare mean for organisational crisis plans written before 2023? Plans written before AI became an operational reality, before information warfare became standard statecraft, and before financial instruments emerged as conflict tools are likely to contain significant assumption gaps. The 2026 convergence event — simultaneous kinetic operations, narrative warfare, insurance-market disruption, and civilian app compromise — is the model for the next crisis, not the exception to it.

Philippe Borremans is a crisis communication specialist with 25 years of experience and the founder of RiskComms. He publishes the Wag the Dog Newsletter, which covers the intersection of crisis communication, narrative intelligence, and emerging threat environments. Related reading: Who Benefits When the Public is Confused? | The GCC SIRTREP