California Lawmakers Amend New AI Bill, Verdict Expected by End of August

California’s State Assembly’s Appropriations Committee accepted amendments to the new bill S.B. 1047 this Thursday, and it could potentially become a law by the end of the month.

The “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act” requires companies working with AI products and services in California to test safety before releasing the technologies. It aims to prevent major AI consequences and guarantee safety.

According to The New York Times, if approved, California’s attorney general could sue companies if these technologies cause harm—like human casualties or mass property damage—including local tech giants like Google, OpenAI, and Meta.

Senator Scott Wiener wrote the bill in February and since then it has gone through multiple amendments, especially after the increasing concerns and requests from multiple tech companies who oppose it. This would be one of the final stages of the process to approve it.

According to TechCrunch, the most recent amendments have been suggested by the AI company Anthropic.

“We accepted a number of very reasonable amendments proposed, and I believe we’ve addressed the core concerns expressed by Anthropic and many others in the industry,” said Senator Wiener to TechCrunch. “These amendments build on significant changes to SB 1047 I made previously to accommodate the unique needs of the open source community, which is an important source of innovation.”

One of the significant changes is that the bill will not punish companies for not failing to apply good safety practices before any major consequence has happened, only after, as recommended by Anthropic.

A spokesperson from Google said to The New York Times that their concerns “still stand,” Anthropic said it will carefully review the updates, and Meta and OpenAI—even after its recent report on GPT4-o’s safety which aligns with the bill’s requirements— declined to comment on the new changes.

Most AI companies are concerned that this bill will discourage the development of the technology or make companies move headquarters to other states. If approved by the Democratic-majority Legislature’s final vote, then the California Senate, and finally the Governor, it could become the first AI safety law in the United States.