California AI Bill: SB 1047 Introduced - Latest Updates

New AI Bill Introduced in California

The primary sponsor of California’s SB 1047, widely considered the most debated AI safety legislation of 2024, has proposed a new bill with the potential to significantly impact Silicon Valley.

State Senator Scott Wiener of California unveiled a new legislative proposal on Friday designed to safeguard employees within prominent AI laboratories. This measure would empower them to voice concerns if they believe their company’s AI systems present a “critical risk” to society. SB 53 also aims to establish a publicly accessible cloud computing resource, termed CalCompute, to provide researchers and startups with the computational power needed to foster AI development for public benefit.

Background on SB 1047

Wiener’s previous AI bill, SB 1047, ignited a national discussion regarding the regulation of large-scale AI systems and their potential for causing significant harm. The intent of SB 1047 was to mitigate the possibility of extremely large AI models triggering catastrophic events, such as substantial financial losses exceeding $500 million or loss of life. Ultimately, however, Governor Gavin Newsom vetoed the bill in September, citing concerns about its overall approach.

The debate surrounding SB 1047 rapidly became contentious. Certain leaders in Silicon Valley argued that the bill would undermine the United States’ competitive position in the global AI landscape. They asserted that the legislation stemmed from unfounded anxieties about AI systems leading to fictional, apocalyptic scenarios. Senator Wiener, in turn, accused some venture capitalists of orchestrating a “propaganda campaign” against his bill, specifically referencing a claim by Y Combinator that SB 1047 could result in imprisonment for startup founders – a claim disputed by experts.

Key Components of SB 53

SB 53 effectively incorporates the less contentious elements of SB 1047, including provisions for whistleblower protection and the creation of the CalCompute cluster, into a revised AI bill.

Wiener is not avoiding the topic of potential existential risks associated with AI in SB 53. The bill explicitly protects whistleblowers who raise concerns about their employers developing AI systems that pose a “critical risk.” This risk is defined as a “foreseeable or material risk that a developer’s development, storage, or deployment of a foundation model, as defined, will result in the death of, or serious injury to, more than 100 people, or more than $1 billion in damage to rights in money or property.”

The bill restricts frontier AI model developers – including companies like OpenAI, Anthropic, and xAI – from retaliating against employees who disclose pertinent information to California’s Attorney General, federal agencies, or fellow employees. These developers would also be obligated to provide feedback to whistleblowers regarding specific internal processes they have flagged as problematic.

CalCompute: A Public Computing Resource

Regarding CalCompute, SB 53 proposes the formation of a working group to oversee the development of a public cloud computing cluster. This group will comprise representatives from the University of California system, alongside researchers from both public and private sectors. They will formulate recommendations concerning the construction of CalCompute, its optimal size, and the criteria for user and organizational access.

Legislative Outlook

It is still early in the legislative process for SB 53. The bill must undergo review and approval by California’s legislative bodies before reaching Governor Newsom for consideration. State legislators will undoubtedly be monitoring the response from Silicon Valley.

However, 2025 may present greater challenges for the passage of AI safety legislation compared to 2024. While California enacted 18 AI-related bills in 2024, momentum appears to have shifted away from the focus on potential AI-related dangers.

Vice President J.D. Vance indicated at the Paris AI Action Summit that the United States is prioritizing AI innovation over AI safety concerns. Although the CalCompute cluster envisioned by SB 53 could be viewed as a catalyst for AI advancement, the future of legislative efforts addressing existential AI risks remains uncertain in 2025.