Anthropic bioweapon use admission fuels AI regulation calls

The call for regulation follows a warning from inside the industry itself: Anthropic's recent admission that "threat actors" used its models to try to develop bioweapons came after a researcher left the firm over fears it was losing control of the technology.

That admission has landed in a policy environment that is far from settled. Frontier labs, including OpenAI's chief, have said they are open to restrictions — even considering slowing development. But policymakers remain divided on how to regulate AI, and the US-China race for AI supremacy sharpens the disagreement. Politico reported that members of Congress are confronting the reality that they are "ill-equipped to rein [AI] in."

The stakes may therefore hinge less on whether Washington acts than on whether it can agree on a mechanism, especially with the US and China racing for AI supremacy. For frontier labs and regulators alike, the open question appears to be whether voluntary openness to restrictions from the labs can translate into a workable policy.