SB 53, the AI safety and transparency legislation recently enacted by California Governor Gavin Newsom, demonstrates that state-level oversight doesn't necessarily impede advancements in artificial intelligence.
This perspective comes from Adam Billen, vice president of public policy at the youth-driven advocacy group Encode AI, who appeared on today’s episode of Equity.
“The truth is, lawmakers recognize the need for action, and their experience with countless other issues has shown that it’s possible to craft laws that truly safeguard innovation—which I care about—while also ensuring these technologies are safe,” Billen explained to TechCrunch.
Essentially, SB 53 stands as the first law of its kind in the country, compelling major AI developers to disclose their safety and security measures—particularly regarding how they mitigate catastrophic threats, such as the misuse of AI for cyberattacks on vital infrastructure or the creation of biological weapons. The statute also requires adherence to these protocols, with enforcement handled by the Office of Emergency Services.
“The requirements in this bill are already being met by many companies,” Billen told TechCrunch. “They conduct safety evaluations on their models and publish model cards. Are some firms beginning to cut corners in certain areas? Yes, and that’s why legislation like this is necessary.”
Billen further pointed out that some AI companies have policies allowing them to relax safety measures when facing competitive pressures. For instance, OpenAI has openly acknowledged it might “adjust” its safety protocols if a competitor launches a high-risk system without comparable safeguards. Billen believes that policy can hold companies accountable to their safety commitments, preventing them from compromising standards due to market or financial pressures.
Although public resistance to SB 53 was less pronounced than it was for its predecessor, SB 1047—which Newsom vetoed last year—many in Silicon Valley and the AI sector continue to argue that almost any regulation of AI will stifle progress and ultimately put the U.S. at a disadvantage in its competition with China.
This is why companies like Meta, venture capitalists such as Andreessen Horowitz, and influential figures like OpenAI president Greg Brockman are investing hundreds of millions in super PACs to support pro-AI candidates in state elections. It also explains their earlier push this year for an AI moratorium that would have blocked states from regulating AI for a decade.
Encode AI led a coalition of over 200 groups to help defeat that proposal, but Billen says the battle is ongoing. Senator Ted Cruz, who advocated for the moratorium, is now pursuing a different approach to achieve federal preemption over state laws. In September, Cruz introduced the SANDBOX Act, which would let AI firms apply for waivers to bypass certain federal rules for up to ten years. Billen also expects a new bill to set a federal AI standard, which would be presented as a compromise but would, in effect, override state regulations.
He cautioned that narrowly tailored federal AI laws could “erase federalism for the most significant technology of our era.”
“If you told me SB 53 was intended to replace all state-level AI laws and address every possible risk, I’d say that’s not a good approach—this bill is tailored to a specific set of issues,” Billen remarked.
Adam Billet, vice president of public policy, Encode AI
Image Credits:Encode AI
While he acknowledges the importance of the AI race with China and the need for regulations that foster American innovation, he argues that eliminating state laws—which primarily address issues like deepfakes, transparency, algorithmic bias, child protection, and government AI use—is not the solution.
“Will bills like SB 53 prevent us from surpassing China? Absolutely not,” he stated. “Claiming otherwise is simply not intellectually honest.”
He continued: “If your main concern is winning the AI race against China—and I share that concern—then you’d be advocating for things like congressional export controls,” Billen said. “You’d ensure American companies have access to the necessary chips. But that’s not what the industry is prioritizing.”
Legislation such as the Chip Security Act aims to stop advanced AI chips from being diverted to China by implementing export controls and tracking, while the CHIPS and Science Act is designed to enhance domestic chip manufacturing. However, some leading tech firms, including OpenAI and Nvidia, have voiced reservations or opposition to parts of these initiatives, citing worries about their effectiveness, competitiveness, and potential security risks.
Nvidia, for example, has strong financial motivations to keep selling chips to China, which has long been a major market for the company. Billen speculated that OpenAI might be hesitant to push for chip export restrictions to maintain good relationships with key suppliers like Nvidia.
There has also been mixed messaging from the Trump administration. Just three months after expanding a ban on advanced AI chip exports to China in April 2025, the administration reversed its stance, allowing Nvidia and AMD to sell certain chips to China in exchange for 15% of the revenue.
“You see lawmakers moving toward proposals like the Chip Security Act to impose export controls on China,” Billen said. “Meanwhile, there’s an ongoing effort to promote the narrative that state-level bills—which are actually quite moderate—should be scrapped.”
Billen also described SB 53 as a testament to democracy in practice—an example of industry and lawmakers collaborating to create legislation that everyone can accept. He admitted the process is “messy and imperfect,” but emphasized that this kind of democratic and federalist process is the bedrock of the nation and its economy, and he hopes it will continue to thrive.
“I believe SB 53 is one of the clearest examples that this approach can still succeed,” he concluded.
