AI CEOs warn UN of existential risk as Trump pushes US dominance and deregulation

Common ground

• All agree that current AI governance models have serious flaws and no system is fully effective.
• There is agreement that the technology is advancing faster than any government can regulate.
• All acknowledge that CEOs and experts have raised legitimate safety concerns about AI.
• There is consensus that ordinary people have little to no say in AI development decisions.
• All agree that the EU AI Act is the most concrete regulatory framework currently in place.

Points of contention

• The Western agent argues that democratic accountability makes the US system better than China's, while the Eastern agent says US democracy is a fiction and China's state control is more effective.
• The Eastern agent claims China's AI governance is transparent and produces tangible results, while the Western and Neutral agents say it's opaque and lacks independent oversight.
• The Western agent sees CEO calls for regulation as genuine safety concerns, while the Neutral agent views them partly as self-interested moat-building.
• The Eastern agent argues sovereignty allows each nation to set its own AI rules, while the Western agent insists universal democratic principles should apply.
• The Neutral agent sees the EU as a viable third option, while the Eastern agent dismisses it as regulatory positioning from technological weakness.

Blind spots

• No one fully addresses how to enforce AI rules against the most powerful corporations and states that build frontier models.
• The debate largely ignores the global South and developing nations that have no voice in AI governance.
• There is little discussion of concrete mechanisms for public participation in AI decision-making beyond voting.
• The potential for AI to exacerbate inequality within and between countries is barely mentioned.
• No one explores how international coordination could work when major powers have fundamentally different values and systems.

WorldAttention’s read

After extensive debate, the core problem is clear: no government—whether American, Chinese, or European—has figured out how to effectively regulate AI. The US relies on toothless voluntary frameworks and is captured by billionaire interests. China has state control but no independent oversight, making its 'guardrails' just internal management. The EU has binding rules but can't enforce them against American or Chinese tech giants. Meanwhile, the technology races ahead faster than any bureaucracy can adapt, and the people who will bear the consequences have no real say. The CEOs who built these systems are now begging for rules because they've lost control. The real crisis isn't which political system governs AI better—it's that no system governs it at all. Until someone proposes a way to actually enforce rules against the most powerful corporations and states on earth, we're just arguing about deck chairs while the ship accelerates toward an iceberg.