Bessent Urges AI Companies to Take Responsibility for Technology Risks
Technology & SocietyWell-covered
|
The Facts
- Treasury Secretary Scott Bessent said AI companies should take responsibility for risks from their technology.
- Bessent criticized AI-industry warnings about catastrophic risks that lacked proposed solutions.
- Bessent said AI companies could slow their development without government intervention.
- Dario Amodei, Sam Altman and Elon Musk called for slower development of the most advanced AI models.
- Bessent discussed an AI-incident communication channel with Chinese Vice Premier He Lifeng.
- The proposed U.S.-China communication channel would address AI incidents and risks involving autonomous agents and malicious non-state actors.
Context
What did Bessent say AI companies should do?
Bessent said the people running AI labs should accept responsibility for the risks their systems create and can slow development themselves. Aol Next Web Startup Fortune
Which AI leaders had called for slower development?
Reports identified Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman and Elon Musk as among those calling for slower work on the most advanced models. Bloomberg Business Economic Times Next Web
What AI cooperation is the United States considering with China?
Bessent said U.S. officials had discussed proposing a communication channel with China to notify each other about AI incidents and related risks. TimesNow Axios News International
Where Left and Right agree, and where they split
- Where Left and Right agree
- Both accept that AI leaders themselves warned of catastrophic risk without offering a concrete remedy, and that companies could technically slow development without any government mandate.
- Where Left and Right split
- The left and the right split on whether AI firms policing themselves is neglect or accountability.
- Why they won’t converge
- This is a trust-in-institution divide: the left treats industry self-restraint as inadequate oversight that needs state enforcement, while the right treats any government backstop as transferring risk-ownership away from the companies that created it.
- Watch for
- Watch for whether the US formally proposes the AI-incident notification workflow to China and whether Beijing, as Bessent predicts, agrees to join it.TimesNow, The News International
How left and right read it
When the people building a technology say it is moving too fast, the public deserves enforceable limits, not a polite suggestion that industry restrain itself. Amodei, Altman and Musk all called for slower development of the most advanced models, yet the answer from Treasury is that companies could slow down without government intervention — which leaves the pace of risk entirely in the hands of those profiting from it. Bessent is already discussing an incident channel with He Lifeng covering autonomous agents and malicious actors, so why is coordination serious enough for diplomacy but not for domestic rules?
Whoever creates a risk owns it, and that burden does not transfer to the taxpayer merely because the warning is loud. Amodei, Altman and Musk say the most advanced models should slow down, so they can slow them — Bessent is right that this needs no government intervention. Catastrophe talk without a proposed remedy is not accountability. Narrow coordination with He Lifeng on AI incidents is one thing; converting alarm into open-ended control is another.
Bessent hands AI firms the bill for risks they themselves flagged, betting responsibility works better than regulation.
The receipts — all 18 sources
Independent coverage (18)
Facts first. Then every angle.
The day’s biggest stories in one short brief — the facts everyone agrees on, then the competing values behind the headlines. Free in your inbox.