California Passes First U.S. AI Safety Law: What It Says and How Industry Is Responding

California Governor Gavin Newsom has signed Senate Bill 53 (SB 53), the first state law in the United States to impose specific transparency and safety requirements on large companies working in the field of artificial intelligence. The bill is already being called a landmark step that could set the tone for future AI regulation in other regions of the country.

The bill, which passed the California legislature two weeks ago, went into effect after the governor signed it. It requires leading artificial intelligence labs, including OpenAI, Anthropic, Meta, and Google DeepMind, to publish data on their safety protocols. The bill also guarantees protection for so-called “whistleblowers” ​​— employees who report violations or unsafe practices at companies.

SB 53 also creates a mechanism for interaction with the California Office of Emergency Services, where companies and the public can report potentially dangerous incidents related to AI. Such cases include cybercrime, fraudulent actions or unpredictable behavior of models. At the same time, the law explicitly requires reporting of incidents, even if they are not provided for by the current European AI Act (EU AI Act).

The reaction of the technology industry to SB 53 was mixed. Anthropic expressed support for the new rules, while Meta and OpenAI opposed them. The latter even published an open letter to the governor urging him not to sign the document. Technology companies warn that similar laws at the state level could create a “regulatory patchwork” that would harm the development of innovation.

Despite this, supporters of the law see SB 53 as an important step in countering the risks associated with the rapid development of technology. Other states are now looking to California. A similar bill has already passed the New York legislature and is now awaiting a decision from Governor Katie Gokul.

Newsom said in a statement that the law strikes the right balance: “California has proven that it can set rules that protect the public while allowing industry to thrive. AI is a new frontier for innovation, and our state is not only ready, but also a national leader in the safety of this technology.”

It is worth noting that this is the second attempt to pass such a regulation. Last year, the governor vetoed a broader bill, SB 1047, due to strong industry opposition. This time, the author of the new law, Senator Scott Wiener, actively consulted with large companies to overcome the most vehement objections. In addition, the California parliament has already passed another document — SB 243, which will regulate the activities of companion chatbots and establish legal liability for operators in case of violations.