Treasury Secretary Scott Bessent said AI labs and the people who work in them must take responsibility for the risks of their systems. He urged the industry to pair continued development with practical preparation, arguing that warnings about danger are not enough without solutions. He also said executives who think they may lose control of increasingly powerful models should slow down.
What did Bessent say about AI risks?
In an interview reported by Axios, Bessent argued that responsibility rests with the people building AI systems. “It is the people at the labs who have to accept responsibility, and I agree with that,” he said.
As an Amazon Associate I earn from qualifying purchases.
He said the industry should prepare for serious risks, while criticizing alarm about those risks when it is not accompanied by ways to address them. The Economic Times reproduced his comment that “this alarmism without solutions by some of the AI community, that’s not leadership.” The point was not that risks can be dismissed, but that identifying them should lead to practical responses.
What does “safe acceleration” mean here?
Bessent used the phrase “safe acceleration” to describe pursuing AI development while preparing for potential harms. He did not present it as a technical standard or a defined regulatory program. In the same interview, he said: “We really need to start worrying more about resilience and defense, too.”
#1 Best Overall
That emphasis broadens the discussion beyond how quickly frontier models advance: resilience and defense concern the ability to withstand or respond to harmful uses and failures. The interview report cited concerns including uncontrolled AI agents and non-state actors using AI for cyber or biological threats; it did not establish how likely those events are.
When did Bessent say AI companies should slow down?
Asked about executives who warn they could lose control of increasingly powerful models, Bessent replied, “Well, then they should slow down.” His statement makes continued development conditional on the builders’ ability to manage the risks they identify; it is not evidence of a specific government order to pause development.
Rank #2
What was proposed for U.S.-China AI incident communication?
Axios reported that Bessent discussed a proposed process for the United States and China to communicate about serious AI incidents. The idea was presented as a proposal, not as an operating notification channel or an agreement already in force.
Free tools Windows power users keep installed
One-click scans. No signup required.
The report also distinguished this proposal from domestic U.S. model reviews, which it described as voluntary at the time. It said the administration reserved the right to intervene if a lab proceeded despite safety concerns. Those reported positions do not establish that a particular review process or intervention has since occurred.
Rank #3
How does this relate to Treasury’s financial-sector AI work?
2024 report recommendations
Treasury’s December 2024 report on AI in financial services summarized responses to its request for information and reported receiving 103 comment letters. Its recommendations included collaboration among government, regulators, and financial firms; analysis of possible regulatory gaps and consumer harm; coordination on risk-management frameworks; information sharing; and compliance review before deployment. These were recommendations within the report’s scope, not binding rules.
Read the Treasury announcement of the 2024 report for its scope and recommendations. The 103 submissions measure responses to Treasury’s request, not the frequency or probability of AI harms.
2026 cybersecurity and risk-management initiative
In February 2026, Treasury announced a separate public-private initiative to develop practical AI cybersecurity and risk-management resources for financial institutions. Its workstreams cover governance, data practices, transparency, fraud, and digital identity, with an emphasis on implementation resources. The Treasury announcement describes this financial-sector initiative; it does not establish that the proposed U.S.-China incident channel is operational.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →What is established—and what is not
Bessent’s comments express a policy position: AI developers should accept responsibility, pair warnings with solutions, and take resilience seriously. The cited materials do not provide an independently substantiated rate or forecast for AI incidents, cyber harm, or biological risk. Nor do they show that the bilateral notification proposal has become an implemented system. The distinction matters: calls for action, report recommendations, and announced work on resources are not the same as binding requirements or proof that a policy mechanism is in force.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
