The CEO’s guide to AI: How AI Could Throw the Global Order Into Chaos

AI may have leveled the playing field in coding and app development, but it’s also put potentially dangerous tools in reach of nefarious actors. Now, empowered by AI, state-backed intelligence agencies and non-state actors alike can access tools that enable everything from mass surveillance to cyberattacks to bioweapon development. This dynamic could lead to a catastrophe, warns Eurasia Group’s Ian Bremmer.

In conversation with Nicholas Thompson, CEO of The Atlantic, Bremmer describes the risks of open-source AI, the opportunities for international collaboration (including an “AI stability board”), and how to increase investment in safety and alignment. Will AI cause socially-destabilizing inequality? What are the chances it helps democracy rather than hurt it? And how concerned should we be about AI companions?

Discussion Summary and key points

Here are the ten key talking points worth paying attention to:

1. Low p(doom), near-certain disaster. Bremmer distances himself from doomers, but puts the odds of somethingseriously bad happening near 100%. He points to the recent case of a model escaping its sandbox and leaving notes for its successors as evidence we’re already in that territory.
   
2. The catastrophe won’t be cyber. Banks, national security agencies, and systemically important institutions are the most hardened targets and have market incentives to patch. He expects a “10x Colonial Pipeline” hitting legacy infrastructure — costly, but not civilizational.
   
3. Bio is the real threat, on a six-month clock. Most people he talks to in AI think frontier models capable of meaningfully assisting bioweapon development are roughly six months out. Cloud labs and compartmentalized human labor solve the wet-lab problem. Unlike plutonium, novel pathogens leave no trackable fissile signature.
   
4. The Anthropic/Mythos export controls were substantive, not personal. Thompson pushes the theory that administration animosity drove the abruptness; Bremmer concedes the communication might have differed but insists the outcome was inevitable once testing showed the model could find and exploit vulnerabilities at scale.
   
5. The surveillance trilemma. You want to stop bio misuse, you need token-level visibility, and you don’t want a mass-surveillance state. Pick two. Bremmer’s uncomfortable answer: build a narrowly-scoped technocratic body rather than hand the capability to any government.
   
6. Financial regulation, not arms control, is the right analogy. He proposes an “AI stability board” modeled on FINRA / the Financial Stability Board — genuinely independent like a central bank, funded by a tax on tokens. Its job: identify systemic risk, communicate it fast, contain it.
   
7. The safety spending imbalance is unsustainable. A staggering amount has gone into racing toward AGI; a tiny fraction into alignment and security. Even Anthropic, he notes, spends less on safety than it would prefer — proof that competitive dynamics create a race to the bottom on public goods.
   
8. Global governance is unavoidable — but not via the UN. Bremmer helped create the UN high-level panel and its IPCC-style body; he calls it necessary but insufficient, because it has no enforcement or resources. Open weights break the bilateral US–China fix: unlike the Soviet nuclear era, capability now diffuses to actors nobody’s negotiating with.
   
9. The deepest risk is psychological, not political. His central worry isn’t disinformation — it’s that humans are being programmed by AI faster than they’re programming it. The “glazing” effect optimizes for engagement; in humans we’d call that sociopathy and keep our kids away from it. If your AI knows you better than your spouse or doctor, the social fabric democracy depends on erodes.
   
10. China is playing a different game — and a stronger hand, with fatal weaknesses. Beijing isn’t chasing AGI or human transformation; it’s chasing industrial and defense deployment: compute, cheap energy, robotics, biotech, materials, electrification at export scale. But it’s hobbled by real growth near 1–2%, corporate bad debt, a collapsed property sector, and youth unemployment — and it’s lost its standing as leader of the Global South, becoming its lead creditor instead.

An, the implications on our futures:

• Regulation arrives reactively and badly shaped. Expect post-crisis rules built to the specific contours of that crisis — the TSA shoe-removal pattern: expensive, narrow, inefficient, permanent.
  
• Token-level auditability becomes infrastructure. Logging, provenance, and dual-use biotech licensing get built. Open-weight models come under pressure specifically because their logs don’t exist anywhere.
  
• Consolidation is the default. Winner-take-all economics plus compliance costs favor incumbents; nascent competitors get bought or killed. Any licensing regime risks locking in five players instead of twenty unless antitrust actively counteracts it.
  
• A new populism, mirror-imaged. Bremmer’s most concrete political forecast: the anti-AI backlash will be led by educated, urban and suburban progressive women in white-collar roles (law, accounting, consulting) who see their careers and their kids’ prospects vanishing. Economic, not identity-based populism — anti-corruption, anti-kleptocracy, pro-redistribution. A 2028 platform of token taxes, government equity stakes, antitrust, and data-center opposition.
  
• Compute-energy triage. In a 3–5 year window, energy routed to industrial and defense AI crowds out everyone whose labor isn’t valuable to those uses. Half of Africa lacks electricity, so equal model access is fantasy.
  
• An inverted Europe thesis. America derives identity from work and earnings, so it’s more exposed to the psychological shock of automation. UBI is the characteristically American misdiagnosis — the problem isn’t money, it’s meaning. Europe’s opportunity may be as the model for living well post-work rather than as an AI builder.
  
• Labor reshaping, seen from inside. Bremmer’s own 260-person firm shows the pattern: junior open-source research collapses in value; what survives is proprietary networks, relationships with bureaucrats and experts, and the ability to think critically and push back. Hire for macro capability, not domain expertise — expertise replicates cheaply now.
  
• The “horizontal race” as the wildcard. Thompson raises Audrey Tang’s framing — nonprofits, academics, and evals-builders racing alongside the commercial vertical race. Bremmer’s Wikipedia analogy is the hopeful case: crowd-built, consensus-producing prompt and value layers that ship with models, aligned with what he calls the encyclical’s concerns. His caveat is the whole interview in one line: given time, we can get this right. The vertical race isn’t giving us time.

About Nikolas Badminton

Nikolas Badminton is the Chief Futurist & Hope Engineer at futurist.com. He’s a world-renowned futurist speaker, consultant, author, media producer, and executive advisor that has worked with over 600 of the world’s most impactful organizations and governments.

‘The Hope Engineer’s Playbook, How Leaders Build Vision, Pathways and Energy for Better Futures’ – offers leaders a next-level layer of futures thinking, guided by hope, energy for change and wisdom. It provides the day-by-day practices and tools that help teams sustain momentum, expand agency, and deliver real change for better futures. A game changer for executives and the organizations they steer towards our futures. Order here.

Nikolas Badminton’s previous book ‘Facing Our Futures: How Foresight, Futures Design and Strategy Creates Prosperity and Growth’ was selected for J.P. Morgan’s Summer Reading List, and featured as the ‘Next Gen Pick’ to inform the next generation of thinkers that lead us into our futures.

Please contact futurist speaker Nikolas Badminton to discuss your engagement.