California state Senator Scott Wiener, the architect behind the nation’s most debated AI safety bill of 2024, is returning with a new legislative effort aimed at reshaping the AI landscape. On February 28, Wiener introduced Senate Bill 53 (SB 53), which seeks to provide whistleblower protections for employees at leading AI labs and establish a state-backed cloud computing cluster known as CalCompute to support AI research and development.
Whistleblower Protections for AI Workers
SB 53 comes with a special provision designed to shield employees who raise concerns about AI systems that could pose a “critical risk” to society. The bill explicitly protects workers who report AI risks to California’s Attorney General, federal authorities, or even their colleagues, ensuring that developers of frontier AI models—such as OpenAI, Anthropic, and xAI—cannot retaliate against whistleblowers.
The legislation defines “critical risk” 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.” This language signals Wiener’s continued focus on the potential catastrophic consequences of AI mismanagement, despite the controversy surrounding his previous bill, Senate Bill 1047 (SB 1047).
The Strategic Recalibration From SB 1047 to SB 53
SB 53 showcases a more streamlined approach compared to its predecessor, SB 1047. SB 1047 sought to regulate all aspects of large AI models—from cyberattacks to nuclear disasters—while setting a threshold of $500 million in damages. While it provoked intense debate about the future of AI regulation, it was ultimately vetoed by Governor Gavin Newsom due to concerns about its overly broad regulatory framework.
The political firestorm surrounding SB 1047 intensified, with Silicon Valley leaders accusing the bill of undermining U.S. competitiveness in the international AI arena. Meanwhile, when Wiener accused venture capitalists of running a “propaganda campaign” against the bill, the battle grew even more heated. One of the most incendiary claims came from Y Combinator, which alleged that SB 1047 would imprison startup founders—a claim that experts quickly dismissed as baseless.
CalCompute: A Public Cloud for AI Research
Another key component of SB 53 is CalCompute, a proposed public cloud computing cluster designed to provide essential computing resources for AI research. The bill proposes creating a working group comprising representatives from the University of California, as well as public and private sector researchers, who will determine the scope, scale, and accessibility of the cluster.
CalCompute aims to increase equality in the field of AI research. It seeks to provide startups and independent researchers with the computing power necessary to develop AI technologies that serve the public interest. This initiative could counterbalance the dominance of private AI labs, which often have exclusive access to advanced computing resources.
The Road Ahead
SB 53 is still in the early stages of the legislative process. The legislation must be approved by California’s legislative bodies before reaching Governor Newsom’s desk. Considering the backlash against SB 1047, state lawmakers will be monitoring Silicon Valley’s response to this new bill.
The year 2025 will bring new challenges for AI safety legislation. Although California passed 18 AI-related bills in 2024, the momentum behind AI risk regulation seems to be slowing.
Notably, Vice President J.D. Vance’s remarks at the Paris AI Action Summit suggest that the federal government is prioritizing AI innovation over regulation. While CalCompute aligns with this pro-innovation stance, SB 53’s whistleblower protections may face opposition from industry leaders who oppose restrictions on AI development.
