California has become the first state in the nation to enact an AI safety law. The Transparency in Frontier Artificial Intelligence Act requires major AI firms to regularly report safety information, keeping government apprised of the guardrails imposed when building models as well as ongoing risks presented. California is home to leading AI companies including OpenAI, Alphabet, Anthropic, Meta Platforms, Nvidia and xAI, which means the law will be something of a national standard, as the rules imposed on those companies will have follow-through effects in all states. The law also bolsters whistle-blower protections for employees of the affected firms.

“AI is the new frontier in innovation, and California is not only here for it, but stands strong as a national leader,” said California Governor Gavin Newsom in a statement that notes the law installs “commonsense guardrails on the development of frontier artificial intelligence models” while balancing the interests continuing development with building the public’s trust. SB 53 was signed into law September 29.

“State Senator Scott Wiener, a Democrat from San Francisco, who proposed the legislation, said the law was crucial to fill a vacuum to protect consumers from potential harms from AI,” writes The New York Times, explaining that “the closely watched California law will escalate the tech industry’s war against states taking regulation of AI into their own hands.”

The “One Big Beautiful Bill” that became federal law on July 4 originally contained a 10-year moratorium on state and local AI laws, but that wildly unpopular provision was stripped during reconciliation by a vote of 99 to 1.

NYT says the big AI firms find compliance with a patchwork of state rules unduly burdensome and have instead been pushing for federal legislation. In June, Texas became the second state to enact comprehensive AI legislation with the Texas Responsible AI Governance Act. The first major state AI law, the Colorado AI Act of May 2024, centers on consumer affairs.

California’s SB 53 is the first sweeping law to target development of the most sophisticated frontier models. New York is said to have taken a similar approach with the Responsible AI Safety and Education Act, or RAISE Act, which made it to Governor Kathy Hochul’s desk on June 12.

“This year, 38 states passed or enacted about 100 AI regulations, according to the National Conference of State Legislatures,” reports NYT.

“If federal standards are put in place, Newsom said, the state legislature should ‘ensure alignment with those standards — all while maintaining the high bar established by SB 53,’” reports CNBC. However, federal law preempts state law.

California’s AI Act is the second such bill to have made it to Newsom’s desk. Last year he vetoed a bill that would have required large AI companies to hire third-party auditors to annually review risk assessments. Politico says companies including Meta and OpenAI that lobbied against the earlier bill “did not publicly oppose Weiner’s bill,” which is considered milder.

Anthropic was the only company that went on record to actively support SB 53.