OpenAI’s Sam Altman said he supports legislation to slow AI development during his visit to Capitol Hill on Wednesday, as debate swirls in Washington and Silicon Valley over safety risks posed by frontier models.
OpenAI changed tack after two of its agents went rogue, days ahead of the White House’s Aug. 1 deadline for developing an AI safety framework; Washington is also considering curbs on China’s open-source AI, generating disagreements between tech leaders over the necessity of guardrails.
Lawmakers are under pressure to move as quickly as AI has, but “making decisions based on fears” risks ineffective or self-defeating policy, Semafor Tech Editor Reed Albergotti argued: Instead of banning Chinese models, Washington’s “best counterpunch” is to help companies use them safely.