Two former OpenAI researchers who resigned this year over safety concerns say they are disappointed but not surprised by OpenAI’s decision to oppose California’s AI disaster prevention bill, SB 1047. Daniel Kokotajlo and William Saunders had previously warned that OpenAI was in a “reckless” race for dominance.
“Sam Altman, our former boss, has repeatedly called for regulation of AI,” they write in a letter obtained by Politico, calling on California Governor Gavin Newsom to sign the bill. “Now that actual regulation is on the table, he is against it.” The two add: “We hope that with the appropriate regulation, OpenAI can still live up to its mission of building AGI safely.”
In a statement to TechCrunch, an OpenAI spokesperson said in response to former employees’ questions that the startup “strongly disagrees with the misrepresentation of our position on SB 1047.” The spokesperson pointed to AI bills in Congress that OpenAI has supported, noting that “safety regulations for breakthrough AI should be implemented at the federal level because they have implications for national security and competitiveness.”
OpenAI competitor Anthropic has expressed support for the bill, raising specific concerns and asking for changes. Several of those changes have since been incorporated, and on Thursday CEO Dario Amodei wrote to Newsom that the “benefits of the current bill likely outweigh the costs,” but without fully supporting the bill.
