California State Senator Scott Wiener has introduced SB 53, a groundbreaking bill that would establish comprehensive transparency standards for major AI companies, including disclosures of safety protocols and incident reports affecting public safety.

If enacted, California would become the first U.S. state to impose such rigorous transparency requirements on large AI developers like OpenAI, Google, xAI, and Anthropic.

The bill mandates that these firms publicly disclose their safety and security practices and report critical safety incidents to the California Attorney General, aiming to boost industry accountability and public trust.

While some industry stakeholders express caution, fearing the legislation could hinder innovation, supporters believe that increased transparency will foster trust and safety in AI development.

California's efforts are part of a broader trend, with states like New York considering similar legislation such as the RAISE Act, especially as federal regulation remains uncertain after a proposed AI moratorium was rejected by the Senate.

Unlike previous legislation like SB 1047, SB 53 does not hold AI developers liable for harms caused by their models, aiming to avoid overburdening startups and researchers working with open-source or fine-tuned models.

Supporters, including AI researchers and industry leaders, see SB 53 as a balanced approach that encourages innovation while implementing safeguards to address potential catastrophic risks posed by advanced AI models.

The bill is a scaled-back version of last year's vetoed SB 1047, shifting focus from legal liability to transparency to promote both public accountability and technological progress.

California is leading the way in AI regulation due to the absence of federal legislation, setting a precedent that could influence other states and national policy, much like its previous leadership with the California Consumer Privacy Act.

The legislation builds on voluntary safety commitments made by leading AI companies like Meta, Google, OpenAI, and Anthropic, formalizing these into law to ensure industry-wide accountability.

Wiener’s previous bill, SB 1047, was vetoed after opposition from Silicon Valley, but SB 53 incorporates recommendations from California’s AI policy advisory group to create a more moderate and industry-friendly framework.

Currently, SB 53 is progressing through the California State Assembly’s Committee on Privacy and Consumer Protection and must clear additional legislative hurdles before reaching Governor Newsom for approval.

The bill emphasizes risk mitigation by requiring companies to report major safety incidents within 15 days, and it includes protections for whistleblowers who disclose critical risks or violations, targeting well-resourced firms working on advanced AI models.