California Governor Newsom signed the Transparency in Frontier Artificial Intelligence Act into law, mandating AI firms with at least $500 million in annual revenue to publish safety protocols on their websites and report safety incidents to state authorities. The measure does not require formal safety testing or a kill switch.
In contrast to the vetoed S.B. 1047, SB53 centers on transparency rather than verification, asking companies to explain how they incorporate national, international, and industry-consensus best practices into AI development without prescribing specific standards or independent checks.
The law imposes civil penalties of up to $1 million per violation for noncompliance and includes whistleblower protections for employees raising safety concerns. It defines catastrophic risk narrowly as incidents potentially causing 50 or more deaths or $1 billion in damage through weapons assistance, autonomous acts, or loss of control.
The signing underscores California’s role as a regulatory testing ground for AI and could set a broader legislative precedent, given the state’s concentration of AI firms and venture capital. Observers say California’s approach may influence how other states—and potentially federal policy—address AI safety and accountability.
While SB53 emphasizes disclosure over mandatory testing, critics warn that the law’s emphasis on voluntary compliance beyond reporting may limit its protective reach, and proponents argue it provides pragmatic guardrails that allow the industry to innovate while increasing visibility into safety practices.
