By Erin Keen, Staff Writer
In the absence of comprehensive federal legislation on Artificial Intelligence (AI), California leads as a frontier on AI safety issues. On September 29, Governor Gavin Newsom signed into law Senate Bill 53, known as the Transparency in Frontier Artificial Intelligence Act. SB 53 is the first piece of legislation creating a regulatory framework for advanced AI models. [1]
Last September, Gov. Newsom vetoed a similar bill due to harsh pushback from Big Tech lobbyists. The industry criticized the bill’s provisions for being too “heavy-handed,” because it included the requirement of protecting whistleblowers and preventing AI from being used for critical harms. [2]
The new law requires frontier developer companies that train AI foundation models to publish and explain safety frameworks, which identify and mitigate catastrophic risks. The statute defines a “catastrophic risk” as a foreseeable, material risk of more than 50 deaths, or greater than $1 billion in property damages. [3]
The Working Group report states “Evidence that foundation models contribute to both chemical, biological, radiological, and nuclear (CBRN) weapons risks for novices and loss of control concerns have grown, even since the release of the draft of this report in March 2025.” [4] While several AI developers have entered voluntary commitments to AI safety, SB 53 ensures greater accountability across the industry. [5]
Under the new bill, the primary goal is to balance AI risks with benefits. Companies are required to report any critical safety incidents within 15 days to California’s Office of Emergency Services, publicly disclose their safety and security protocols, and protect whistleblowers. [6] For any violations, the Attorney General may impose civil liability. However, SB 53 does not impose any new liability for harm caused by AI systems. [7]
The need for regulations is heightened as AI begins to become commonly deployed in consequential areas. In one instance, a language model from the notable AI company Anthropic threatened to blackmail engineers by threatening to reveal affairs if the system were taken offline. [8] In a separate assessment, models were able to tell when they were being evaluated, corrupting the purpose of the assessments. [9]
Both critics and supporters of the bill have voiced their opinions on the matter. Anthropic backed the bill, with the co-founder sharing that the bill “establishes meaningful transparency requirements for frontier AI companies without imposing prescriptive technical mandates.” [10]
Some companies believe state‑by‑state rules could create a patchwork that complicates compliance and have urged Congress to set national standards. [11] Furthermore, Big Tech firms such as Apple and Amazon state that SB 53’s enactment would create “sweeping restrictions on AI developers.” [12]
California currently dominates and leads the U.S. in the AI sector. California is also home to three of the four largest tech companies, collectively surpassing the $3 trillion valuation mark. [13] Officials in California view the Bill as part of a broader strategy to lead the rest of the country in responsible AI use, while federal legislation remains uncertain. [14]
Whether or not SB 53 becomes part of national legislation will depend on how effectively the Bill balances different interests while allowing California to continue leading in the tech sector. With the bill going into effect sometime in early 2026, published anonymized annual summaries will begin being published in 2027. [15]
[1] https://www.gov.ca.gov/2025/09/29/governor-newsom-signs-sb-53-advancing-californias-world-leading-artificial-intelligence-industry/
[2] https://www.sacbee.com/news/politics-government/capitol-alert/article312274887.html
[3] https://www.lawcommentary.com/articles/california-enacts-ai-safety-disclosure-law-targeting-frontier-model-developers
[4] https://sd11.senate.ca.gov/news/governor-newsom-signs-senator-wieners-landmark-ai-law-set-commonsense-guardrails-boost
[8] https://calmatters.org/newsletter/artificial-intelligence-regulations/
[10] https://www.sacbee.com/news/politics-government/capitol-alert/article312274887.html;
[11]https://www.lawcommentary.com/articles/california-enacts-ai-safety-disclosure-law-targeting-frontier-model-developers
[12] https://www.sacbee.com/news/politics-government/capitol-alert/article312274887.html
[13] https://www.gov.ca.gov/2025/09/29/governor-newsom-signs-sb-53-advancing-californias-world-leading-artificial-intelligence-industry/
[14] https://www.lawcommentary.com/articles/california-enacts-ai-safety-disclosure-law-targeting-frontier-model-developers