California Sets New Standards for Frontier AI Safety With SB 53

California’s adoption of SB 53 marks the introduction of the first statewide requirements in the U.S. for transparency and risk management in frontier artificial intelligence development.

The Transparency in Frontier Artificial Intelligence Act, authored by State Senator Scott Wiener, was signed into law in 2025 and took effect on January 1, 2026. The law mandates that developers of advanced AI systems publish frameworks detailing how they assess and manage catastrophic risks.

SB 53 defines catastrophic risk as a foreseeable and material risk from a frontier AI model that could result in death or serious injury to more than 50 people, or cause property damage exceeding $1 billion in a single incident. The legislation also includes whistleblower protections for employees who report critical risks or false statements about risk management related to AI activities.

Anthropic researchers publicly warned that superintelligent AI could pose existential threats, including the potential to destroy humanity. In response to these concerns, Anthropic endorsed SB 53 and released a compliance framework outlining its approach to managing catastrophic risks as required by the law.

What the numbers show

• SB 53 took effect on January 1, 2026.
• Catastrophic risk is defined as harm to over 50 people or more than $1 billion in damages.
• SB 53 is the first U.S. law to require transparency on frontier AI risks.

Senator Wiener, together with lawmakers from New York and Illinois, called on leading AI companies to negotiate a joint, independently verified framework to pace the development of frontier AI until safety research can address emerging risks. This proposal seeks to align industry practices with evolving safety standards.

SB 53 received support from a coalition that included Encode Justice, Secure AI Project, Economic Security California Action, civil society organizations, labor groups, AI safety organizations, and Anthropic. This broad backing reflects coordinated efforts among various groups to address AI-related risks.

The law’s whistleblower provisions protect employees who report that their employer’s AI activities pose critical risks or involve misleading statements about risk management. These measures are intended to encourage transparency and accountability within organizations developing advanced AI systems.

Industry reaction

Anthropic supported the passage of SB 53 and published a compliance framework to meet the law’s requirements. The company stated that its framework explains how it assesses and manages catastrophic risks in accordance with the new legislation.

Other organizations, including civil society groups and labor organizations, joined the coalition backing SB 53. Their support was documented in legislative analyses and public statements during the bill’s passage.

* This article is based on publicly available information at the time of writing.