Altman says OpenAI and Anthropic still split on AI rules

Asked where he parts ways with Anthropic and its chief Dario Amodei, Altman said the world should “accept some bad things happening” in return for the benefits and for people’s agency. He rejected the idea of one San Francisco lab holding the most powerful systems and handing out the gains, and said he would not swap broad access for a promise of zero hacks or scams. His limit is the catastrophic end of the scale, including AI slipping out of human control.

The distance between the two labs has shrunk. SiliconANGLE notes that Altman backed Amodei’s call last month to slow frontier development. Business Insider adds that OpenAI now endorses stricter state safety laws and a bipartisan House bill that would put outside safety evaluators inside the top AI companies, shifts that followed a run of hacking incidents by OpenAI’s agents.

For teams that buy from both labs, the interview spells out the remaining difference: OpenAI argues for wide access with some harm accepted, and Anthropic, in Altman’s account, for tighter control of the most capable systems.