# Bessent: AI labs must bear responsibility for risks

> Treasury Secretary Scott Bessent says hyperscalers cannot escape liability for AI safety failures, rejecting calls for government protection.
- Title: Bessent: AI labs must bear responsibility for risks · Fortune
- Summary: Treasury Secretary Scott Bessent says hyperscalers cannot escape liability for AI safety failures, rejecting calls for government protection. Treasury…
- Keywords: scott bessent, ai safety, hyperscalers, liability, openai, anthropic, regulation, government oversight, Bessent, labs, must, bear
- Source: Fortune — https://fortune.com/2026/09/22/bessent-ai-labs-responsibility-government-liability-shield
- Author: Eleanor Pringle
- Read time: 1 min
- Topics: scott bessent, ai safety, hyperscalers, liability, openai, anthropic, regulation, government oversight
## Researchers warn of extinction risk
Former OpenAI and Anthropic researchers have raised alarm about AI safety. Evan Hubinger at Anthropic warned there is a maximum ten percent chance AI could wipe out humanity within the next decade.

> low but maximum 10% chance AI could wipe out humanity within the next decade
## Bessent rejects liability shield for labs
Treasury Secretary Scott Bessent told CNBC that AI companies cannot escape responsibility for their technology. He said government will not become a liability shield for hyperscalers, no matter what risks they claim.

> We cannot absolve you of responsibility
## Bessent: labs can slow down anytime
Bessent clarified the administration's position: AI labs need to take responsibility for themselves and can slow down their progress whenever they choose. Humans, not AI, are responsible for the technology's risks.

> They can slow down any time they want to
## Trump takes different approach to regulation
President Donald Trump claimed the U.S. already has tremendous regulatory power over AI companies. He said the only guardrails AI needs is a strong and smart president, and criticized Anthropic cofounder Dario Amodei.

> the only control or 'guardrails' that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT
## Amodei and Altman call for government involvement
Anthropic cofounder Dario Amodei wrote a letter calling on labs to slow progress and suggesting U.S. government support would be needed to globally coordinate and legally bind safety standards. OpenAI CEO Sam Altman echoed this argument.

> It can simultaneously be true that... there might be significant risk
## Sanders: binding rules needed, not voluntary standards
Senator Bernie Sanders pointed out that competition and self-regulation are not traditional bedfellows. He argued that when humanity's future is at stake, binding international safety rules are needed, not voluntary industry standards.

> When the future of humanity is at stake, we need binding international safety rules, not voluntary standards from the industry
## Key takeaway

Treasury Secretary Bessent says AI companies must take responsibility for their technology's risks and cannot shift liability to government.