Newsom Faces Decision on California AI Bill

California Governor Gavin Newsom is facing a crucial decision on SB 1047, a controversial bill designed to prevent potential AI disasters. The bill has garnered significant attention and divided the tech industry, with some companies supporting it while others vehemently oppose it. As the bill awaits Newsom's signature, the future of AI regulation in California and potentially the entire nation hangs in the balance.

Key Provisions of SB 1047

SB 1047, authored by state senator Scott Wiener, seeks to prevent catastrophic events caused by large AI models. The bill focuses on mitigating risks such as loss of life and cyberattacks exceeding $500 million in damages.

It establishes liability for AI developers, making them accountable for potential harms caused by their models.

It grants California's attorney general the authority to sue AI companies for significant penalties if their technology is involved in a catastrophic event.

The bill mandates that AI models subject to the regulations must have a "kill switch" allowing for their immediate shutdown if deemed dangerous.

Why Newsom Might Sign SB 1047

Proponents of SB 1047 argue that it is necessary to hold Big Tech accountable and prevent potential AI-related disasters. They emphasize the need for greater liability and regulation in the tech industry, drawing comparisons to past failures in regulating technologies like firearms.

Senator Wiener believes that Silicon Valley needs more accountability and that lessons must be learned from past mistakes in tech regulation.

Some AI executives, including Elon Musk and Sophia Velastegui, have expressed cautious optimism about SB 1047, suggesting it represents a good compromise.

Anthropic, an AI startup, has supported certain aspects of the bill and believes its benefits outweigh the potential costs.

Why Newsom Might Veto SB 1047

Opponents of SB 1047 argue that it could stifle innovation and have a chilling effect on California's AI industry. They believe the bill's liability provisions could create unnecessary burdens and discourage companies from developing and deploying AI technologies in the state.

The tech industry, including OpenAI, a16z, and Speaker Nancy Pelosi, has voiced strong opposition to the bill, fearing it will create an unfavorable regulatory environment for AI development.

They argue that SB 1047 shifts liability in a way that is unprecedented and could discourage investment in AI research and development.

The U.S. Chamber of Commerce has also urged Newsom to veto the bill, citing the importance of AI to the American economy.

If SB 1047 Becomes Law

If Newsom signs the bill into law, it would set a precedent for AI regulation in the United States. The bill's implementation would unfold over several years, with a phased approach to its various provisions.

By January 1, 2025, tech companies would be required to submit safety reports for their AI models.

In 2026, a Board of Frontier Models would be established to review safety reports and make recommendations to the attorney general.

Also in 2026, AI developers would be required to hire auditors to assess their safety practices.

By 2027, the Board of Frontier Models would begin issuing guidance on safe and secure AI model training and operation.

If SB 1047 Gets Vetoed

If Newsom vetoes SB 1047, it is likely that federal regulators would take the lead in regulating AI models. This would likely involve a slower and less stringent regulatory process compared to California's approach.

OpenAI and Anthropic have indicated their willingness to collaborate with the AI Safety Institute, a federal body, on AI safety research.

OpenAI has endorsed a bill that would grant the AI Safety Institute the authority to set standards for AI models.