Former employees unveil OpenAI’s opposition to SB 1047 legislation

Former OpenAI workers say the company should have supported California’s SB 1047 instead of fighting it, and they say AI safety rules are needed even though OpenAI is worried about national security.

Two researchers who used to work for OpenAI but quit this year because they were worried about safety say they are unhappy but not surprised that OpenAI has decided to fight California‘s SB 1047 bill to stop AI disasters.

William Saunders and Daniel Kokotajlo said that OpenAI is racing to be the best without thinking things through.

In a letter shared with Politico, they say, “Our former boss, Sam Altman, has repeatedly called for AI regulation.” The letter also asks California Governor Gavin Newsom to sign the bill. Now, when real rules are being talked about, he is against them. They also say, “We hope that OpenAI can still live up to its mission statement of building AGI safely with the right regulations.”

A spokesperson for OpenAI told the former employees in a statement that the company strongly disagrees with how our opinion on SB 1047 has been portrayed. A spokesperson for OpenAI pointed to AI bills in Congress and said that new AI safety rules should be put in place at the government level because they affect national security and competitiveness.

Anthropic, a company that competes with OpenAI, has said it supports the bill but has some worries and wants changes made. Since then, many of them have been added, and on Thursday, CEO Dario Amodei wrote to Newsom that the current bill’s pros probably beat its cons, but he didn’t fully support it.