Scott Bessent Critiques Pro-Regulation AI CEOs

Scott Bessent, Treasury Secretary, rejects the idea that AI companies should wait for public safeguards. He calls on industry leaders to self-regulate, citing recent decisions at OpenAI and praising the cooperation of laboratories. These statements come as leaders call for government intervention and the White House emphasizes voluntary commitments.

OpenAI's Paused Model and Security Executive's Resignation

Scott Bessent highlighted OpenAI's recent decision to pause a new model after internal testing raised questions about the AI's ability to follow user instructions. He noted that the laboratories have been "very cooperative." Meanwhile, concerns persist within the sector. Sarah Heck, head of public policy at Anthropic, stated on X last week: "You can't guarantee safety from the second position." At OpenAI, David Robinson, head of security, announced his resignation on Saturday, explaining that he believes OpenAI is moving too quickly to ensure its models won't wreak havoc on society.

Bessent Calls for Self-Moderation Rather Than Public Safeguards

Scott Bessent criticized the stance of some AI leaders who favor government-imposed safeguards. During his appearance on The Axios Show on Saturday, he stated: "This alarmism without solutions from some members of the AI community — that's not leadership." He compared these leaders to Hannibal Lecter, saying, "Stop me before I kill again," to illustrate what he sees as the inconsistency of calls for public regulation coming from those developing the technology. According to Bessent, the industry can slow its own development without government intervention.

Between Calls for Intervention and the White House's Stance

Last month, AI leaders called for an organized slowdown of companies in the sector, fearing that the technology could become uncontrollable. Dario Amodei, head of Anthropic, reached out to U.S. authorities to seek regulation of the sector. At the White House, President Donald Trump invited AI leaders to a meeting that resulted in an agreement for regulation on a voluntary basis. Scott Bessent noted that AI companies seem to be taking the executive branch's recommendations into account.