Breaking

Bessent Pushes Back on AI Extinction Warnings While Seeking U.S.-China Incident Channel

October 3, 2026 at 14:15 EDT

Policy & Regulation
Foundation Models
AI Agents

U.S. Treasury Secretary Scott Bessent criticized prominent AI figures' warnings about existential risk as alarmist and unhelpful, while arguing that AI labs should remain responsible for the systems they build. His comments add a sharper government line to a widening dispute over whether frontier AI development should slow down or accelerate under tighter resilience planning.

Oct. 3, 2026 · U.S. Treasury

Bessent Reframes AI Risk

The message is not that advanced AI is harmless. It is that labs should own the risk, governments should build incident channels, and defenses should catch up with frontier-model speed.

80-90%

Estimated capability of some Chinese open models relative to U.S. models

Proposed emergency notification channel with China

Treasury AI roundtables on governance, cybersecurity and financial stability

Capability Gap Narrows

Bessent described some Chinese open models as 80% to 90% as powerful as U.S. systems.

100%U.S. baseline

80%Lower estimate

90%Upper estimate

Risk Response Playbook

Detect

Spot serious agent, cyber or bio misuse signals.

→

Notify

Use a direct government incident channel.

→

Defend

Harden critical systems before incidents scale.

Acceleration Case

Backers of faster deployment emphasize U.S. competitiveness, useful applications and direct liability for companies that ship unsafe systems.

Safety Case

Safety advocates argue voluntary review is too thin for autonomous systems that may affect cyber, biosecurity and critical infrastructure.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.