On August 29, 2024, California's legislature passed a significant bill aimed at regulating advanced artificial intelligence models, despite concerns that it might stifle innovation in the rapidly evolving field. The bill, known as SB 1047, was sponsored by Democratic state senator Scott Wiener of San Francisco, who expressed pride in the diverse coalition backing the measure, which emphasizes both innovation and safety.
Critics, including some Democratic members of Congress, voiced strong opposition, arguing that imposing punitive measures on developers in such an emerging sector could hinder progress. However, the bill gained unexpected support from Elon Musk, who stated that the potential risks of AI justify the need for regulation, acknowledging the controversy it may create.
The proposed legislation, titled the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act, now awaits the signature of Governor Gavin Newsom, whose stance on the matter remains uncertain. He has until September 30 to either sign or veto the bill.
Dan Hendrycks, director of the Center for AI Safety, endorsed the bill, highlighting it as a viable approach to enforce essential safeguards against significant AI risks. Key requirements for developers of large AI models include conducting pre-deployment testing, simulating cybersecurity threats, implementing protective measures, and ensuring whistleblower protections.
To facilitate the bill's approval in Silicon Valley, lawmakers modified it to replace criminal penalties with civil fines for violations. However, opposition persists from influential figures, including Congresswoman Nancy Pelosi, who labeled the bill as well-meaning but misguided.
Wiener contends that AI safety and innovation can coexist, asserting that revisions have addressed many critics' concerns. OpenAI, the organization behind ChatGPT, also opposed the bill, favoring a unified national approach to AI regulations rather than a fragmented system across states. Currently, over 40 states have introduced similar legislation in 2024, with several enacting laws or resolutions to regulate AI.