On Monday, California Gov. Gavin Newsom (D) signed the Transparency in Frontier Artificial Intelligence Act into law, formally designated as SB53. The measure concludes a fraught period of back-and-forth negotiations regarding artificial intelligence oversight within the state and introduces among the most substantial regulatory requirements across the country.

The legislation has drawn intense scrutiny both within California and at the federal level, positioning itself as a pivotal moment in the national conversation around AI governance. The law could establish a precedent that other states adopt—assuming Congress does not supersede state-level requirements with federal standards.

With a technology as transformative as AI, we have a responsibility to support that innovation while putting in place commonsense guardrails to understand and reduce risk. With this law, California is stepping up, once again, as a global leader on both technology innovation and safety.

California State Sen. Scott Wiener (D), who introduced the bill

The Road to Passage

Though SB53 was initially filed in January, its serious movement toward enactment began the previous year when Wiener put forward a more ambitious predecessor, SB1047, titled the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act. In its original form, SB1047 would have mandated that developers of the most potent AI systems perform safety evaluations before deployment and establish a "positive safety determination" confirming the absence of "hazardous capabilities."

Had it passed, SB1047 would have represented the nation's most demanding AI safety framework. The proposal rapidly became contentious, triggering vigorous opposition from technology sector organizations that contended such restrictions would hamper growth in the emerging AI market. Prominent California political figures, including former House Speaker Nancy Pelosi, voiced similar objections, weakening the bill's viability.

Despite winning substantial backing in the state legislature, Newsom rejected SB1047 in September. In his veto message, he argued that the bill "could give the public a false sense of security about controlling this fast-moving technology" because it concentrated solely on the most sophisticated AI systems. The governor emphasized instead that California required "a delicate balance."

To move discussions forward, Newsom assembled the Joint California AI Policy Working Group, tasked with producing a report containing guidance on AI policy. When Wiener presented SB53 in February as SB1047's successor, he noted in his statement that he was "closely monitoring the work of the Governor's AI Working Group."

The working group delivered its final report in June, recommending "targeted interventions" that would balance the technology's "benefits and material harms." Following this, Wiener's team revised SB53 to "more closely align" with the report's findings. The revised version passed through California's legislature in September and received Newsom's signature shortly thereafter.

California has proven that we can establish regulations to protect our communities while also ensuring that the growing AI industry continues to thrive. This legislation strikes that balance.

Gov. Newsom

What SB53 Requires

Developers of the most sophisticated "large frontier" models must post a publicly accessible framework on their websites detailing their approach to various safety matters. This framework must cover standards and practices, methods for identifying "catastrophic" risks, and procedures for managing "critical safety incidents." The law establishes a computational benchmark for identifying frontier models and designates "large" frontier developers as those with more than half a billion dollars in yearly revenue.

Companies must also distribute transparency reports either before or at the time they release a "new frontier model" or an updated iteration of an existing model. These reports must explain how the company will maintain its stated framework. The law additionally mandates regular submissions to California's Office of Emergency Services containing "a summary of any assessment of catastrophic risk."

The measure streamlines the process for both ordinary citizens and internal company informants to flag potential safety concerns. It directs California's Office of Emergency Services to establish a system allowing the public to report critical safety incidents. The law also bars companies from implementing policies that restrict, punish, or retaliate against workers for sharing information they have "reasonable cause" to suspect demonstrates that a developer creates "specific and substantial danger to the public health or safety resulting from a catastrophic risk."

Furthermore, SB53 mandates the establishment of a consortium responsible for constructing a "cloud computing cluster" referred to as "CalCompute." This infrastructure is intended to facilitate the "development and deployment of artificial intelligence that is safe, ethical, equitable, and sustainable."

Industry and Advocacy Response

SB53 has garnered backing from a wider spectrum of stakeholders than its predecessor, drawing support from consumer protection organizations alongside certain technology firms.

SB 53's passage marks a notable win for California and the AI industry as a whole. Its adaptability and flexible framework will be essential as AI progresses.

Sunny Gandhi, Encode AI

a key victory for the growing movement in California and across the country to hold Big Tech CEOs accountable for their products, apply basic guardrails to the development and deployment of AI, and protect whistleblowers' ability to step forward.

Sacha Haworth, executive director of Tech Oversight California

Reactions from the technology sector have been considerably more nuanced compared to the response to SB1047. Anthropic, which develops the Claude AI assistant, publicly endorsed the bill in recent weeks. Meta characterized the law as "a positive step" toward "balanced AI regulation," while OpenAI stated it was "pleased to see that California has created a critical path toward harmonization with the federal government," though neither company openly promoted the measure, according to Politico.

Conversely, industry associations that include these companies among their members, such as CCIA and the Chamber of Progress, mounted opposition to the bill and urged Newsom to reject it.

includes some thoughtful provisions that account for the distinct needs of startups … it misses an important mark by regulating how the technology is developed — a move that risks squeezing out startups, slowing innovation, and entrenching the biggest players.

Collin McCune, head of government affairs at Andreessen Horowitz

McCune also hinted at the possibility of federal preemption, stating, "Smart AI regulation can help us win the AI race, but we need the federal government to lead in governing the national AI market. Recent conversations with the House and Senate about a federal AI standard are an encouraging step forward. More to come."

Source: Tech Policy Press