The threat of human extinction will get Congress to act on AI safety...right?

A report examining U.S. AI safety policy says rising existential alarm among some AI researchers has pushed lawmakers and media to consider stricter controls, including “kill switches,” federal standards for safe research, and even pausing development of frontier models. But the article notes a stable national approach would likely require Congressional action, which is unlikely during the current session. It cites Harvard Kennedy School computer scientist Stephen Casper, who said there’s no public evidence that leading companies could fully shut down frontier models in an emergency. Several bills are pending, including the AI Kill Switch Act and the broader FRONTIER Act, while California and other states have already passed measures requiring safety plans, and, in Illinois, third-party audits starting in 2028.