Two former OpenAI researchers who resigned this yr over security considerations say they’re disillusioned but not stunned by OpenAI’s choice to oppose California’s invoice to forestall AI disasters, SB 1047. Daniel Kokotajlo and William Saunders beforehand warned that OpenAI is in a “reckless” race for dominance.
“Sam Altman, our former boss, has repeatedly called for AI regulation,” they write in a letter that was shared with Politico and which urges California Governor Gavin Newsom to signal the invoice. “Now, when actual regulation is on the table, he opposes it.” The two add that, “With appropriate regulation, we hope OpenAI may yet live up to its mission statement of building AGI safely.”
Responding to the previous employees, an OpenAI spokesperson stated the startup, “strongly disagrees with the mischaracterization of our position on SB 1047,” in a press release to Ztoog. The spokesperson pointed to AI payments in Congress OpenAI has endorsed, noting that “frontier AI safety regulations should be implemented at the federal level because of their implications for national security and competitiveness.”
OpenAI rival Anthropic has expressed assist for the invoice whereas presenting particular considerations and asking for amendments. Several have since been included, and on Thursday, CEO Dario Amodei wrote to Newsom, saying the present invoice’s “benefits likely outweigh its costs,” whereas not absolutely endorsing the invoice.