California State Senator Scott Wiener has proposed amendments to Senate Bill 53, aiming to enforce transparency among leading AI developers. The revised bill would require AI companies to publish safety protocols and incident reports. This move follows Governor Newsom's veto of a similar bill last year, shifting the focus from liability to transparency.
SB 53 mandates that companies like OpenAI, Google, and Anthropic disclose their safety and security measures, along with evaluations of potential risks. Critical safety incidents, such as data breaches, must be reported to the California Attorney General. The bill also introduces whistleblower protections for those reporting risks within AI organisations. Additionally, SB 53 proposes CalCompute, a public cloud computing cluster to support AI startups and researchers.
If passed, California would be the first state to impose such transparency requirements, potentially setting a national precedent for AI regulation. The bill seeks to strike a balance between public accountability and fostering innovation in the AI sector.
