California’s AI Safety Law SB 53 Balances Innovation and Regulation, Experts Say

California has taken a groundbreaking step in artificial intelligence oversight with the signing of SB 53, a landmark AI safety and transparency bill. Governor Gavin Newsom approved the legislation this week, positioning California as the first state to mandate safety and disclosure standards for large AI developers. Supporters say the move proves that AI regulation and innovation can coexist, countering Silicon Valley’s long-standing belief that any form of government oversight stifles progress.

According to Adam Billen, vice president of public policy at Encode AI, the law is an example of “smart regulation” that protects both consumers and the tech industry. Speaking on TechCrunch’s “Equity” podcast, Billen explained that SB 53 requires major AI companies to publicly disclose their safety and security protocols, including measures to prevent misuse of AI systems in cyberattacks or biological threats. The law also mandates compliance monitoring by the California Office of Emergency Services, ensuring that companies follow through on their commitments.

Billen noted that while most AI firms already conduct internal safety testing and publish model transparency reports, competitive pressure sometimes tempts them to cut corners. He emphasized that the bill simply reinforces existing ethical standards and prevents companies from backtracking on safety commitments. “Are they starting to skimp in some areas? Yes,” he said. “That’s why bills like this are important.” He added that some industry players, including OpenAI, have admitted they might relax safety measures to keep pace with rivals, a risky trend SB 53 aims to curb.

The bill’s passage comes amid broader political efforts to curb state-level AI regulation. Powerful industry figures and venture capital firms, including Meta, Andreessen Horowitz, and OpenAI president Greg Brockman, have poured millions into political campaigns favoring looser AI laws. Meanwhile, Senator Ted Cruz is pushing federal legislation, such as the SANDBOX Act, that would allow AI companies to bypass regulations for up to a decade. Billen and his coalition at Encode AI, which represents over 200 organizations, argue that these moves threaten state autonomy and could undermine safety measures across the U.S.

Despite industry resistance, Billen believes SB 53 is a model for collaborative democracy — proof that government, advocacy groups, and tech innovators can find common ground. He dismissed claims that state laws like SB 53 would slow America’s race against China in AI, calling such arguments “intellectually dishonest.” Instead, he urged policymakers to focus on export controls and chip manufacturing through initiatives like the Chip Security Act and CHIPS and Science Act, which strengthen U.S. competitiveness without sacrificing accountability. “SB 53 shows democracy still works,” Billen said. “It’s messy, but it’s how progress happens.”

source: techcrunch