California has become the first state to establish a comprehensive framework for AI safety regulation with the signing of SB 53 into law by Governor Gavin Newsom. The legislation requires the largest artificial intelligence companies—such as OpenAI, Anthropic, and others—to publicly disclose their safety protocols and adhere to them, marking a major step toward accountability in the rapidly evolving AI industry. The law is a significant evolution from its predecessor, SB 1047, which faced fierce opposition from major tech companies and was ultimately vetoed by Newsom last year. Unlike the earlier bill, SB 53 was crafted with greater input from industry stakeholders and includes provisions designed to balance innovation with oversight. It focuses specifically on AI systems deemed to pose “high risk” based on their scale, capabilities, and potential societal impact. Under SB 53, companies must conduct and report on safety evaluations before deploying advanced AI models, including those with capabilities that could lead to significant harm—such as deepfakes, autonomous weapons, or systems capable of manipulating public opinion. The law also mandates that companies implement and maintain robust risk mitigation strategies and provide transparency about model training data, performance benchmarks, and known limitations. Adam Billen, vice president of public policy at Encode AI, joined TechCrunch’s Equity podcast to explain the significance of the new law. He noted that SB 53’s passage was possible because it avoided overly prescriptive mandates and instead focused on transparency and accountability. “This isn’t about stopping innovation—it’s about making sure companies are responsible for the risks they’re creating,” Billen said. The law also creates a new oversight mechanism: a state AI safety office that will review compliance and have the authority to enforce penalties for noncompliance. While the exact enforcement details are still being developed, the law signals California’s intent to lead the nation in shaping responsible AI governance. The move has already sparked discussion across the tech sector, with some companies welcoming the clarity and predictability the law brings, while others remain concerned about the potential for regulatory fragmentation if multiple states adopt similar rules. Still, experts believe SB 53 could serve as a model for future legislation, not just in the U.S., but globally. As AI continues to reshape industries and daily life, California’s new law represents a pivotal moment in the effort to ensure that innovation does not outpace safety. With other states watching closely, the impact of SB 53 could extend far beyond the Golden State.