California’s AI Safety Bill Passed

California lawmakers passed a contentious artificial-intelligence safety bill on Wednesday. This legislation, known as SB 1047, will undergo one more process vote before it reaches Governor Gavin Newsom, who has until September 30 to decide on its fate.

Industry Reactions

Tech companies involved in generative AI, which can create text, images, or audio from prompts, have expressed significant concerns about the bill. They argue that it could drive AI companies out of California and stifle innovation.

Support and Opposition

Some Democrats in Congress, including Representative Nancy Pelosi, have voiced their opposition to the bill. However, proponents like Tesla CEO Elon Musk, who also leads an AI firm called xAI, have shown support.

Key Provisions of the Bill

The bill mandates safety testing for advanced AI models that cost over $100 million to develop or require substantial computing power. Developers must also establish methods for disabling AI models if they malfunction, effectively creating a ‘kill switch.’

Furthermore, the state attorney general is granted the authority to sue non-compliant developers, especially if there is an ongoing threat, such as AI compromising government systems.

Auditing and Whistleblower Protections

Developers are required to hire third-party auditors to evaluate their safety practices and to implement protections for whistleblowers who report AI abuses.

Legislative Intent

Senator Scott Wiener, the bill’s author and a Democrat representing San Francisco, argues that this legislation is crucial for public safety as AI technology advances.

Industry Concerns

Martin Casado from Andreessen Horowitz expressed hope that Newsom would veto the bill, citing unprecedented bipartisan opposition. Major companies like Google, Microsoft-backed OpenAI, and Meta Platforms have raised concerns in letters to Wiener. Meanwhile, Amazon-backed Anthropic believes the benefits of the bill may outweigh its drawbacks, although some aspects remain troubling.