Anthropic endorses California’s AI safety bill, SB 53

Anthropic has notably endorsed California's SB 53, a bill aiming to impose first-in-the-nation transparency and safety requirements on large AI model developers, specifically targeting 'catastrophic risks' like bioweapons and cyberattacks. This endorsement is significant as it counters broad tech industry opposition, including from groups like CTA and investors such as Andreessen Horowitz, who typically advocate for federal rather than state-level AI regulation due to concerns about stifling innovation. Despite facing potential hurdles and a past gubernatorial veto of a similar bill, SB 53's more modest approach compared to prior legislation suggests a higher likelihood of passage, potentially setting a precedent for state-led AI governance amidst ongoing debates over regulatory authority.

Analysis

Anthropic's endorsement of California's SB 53 bill marks a significant divergence from the broader technology sector's unified opposition to state-level AI regulation. The bill mandates transparency and safety frameworks for frontier AI models from developers with over $500 million in gross revenue, including firms like Google, OpenAI, and Anthropic itself. While industry groups and venture capital firms such as Andreessen Horowitz advocate for a federal-first approach to avoid a patchwork of state laws and maintain innovation pace against China, Anthropic argues that the urgency of AI safety necessitates immediate action. SB 53 is viewed as a more modest and technically grounded proposal than its predecessor, SB 1047, which was vetoed by Governor Gavin Newsom. Specifically, it focuses on preventing 'catastrophic risks'—defined as causing at least 50 deaths or over $1 billion in damages—and has been amended to remove a contentious third-party audit requirement. This legislative restraint has led some policy experts to assess its chances of becoming law as higher, although Governor Newsom's stance remains undeclared, creating a pivotal moment for the future of AI governance in the United States.

Key Decisions for Investors

• Monitor the legislative progress of SB 53 in California, as its passage would set a major precedent for state-level AI regulation, potentially leading to increased compliance costs and operational complexity for large-cap technology companies like Google.
• Assess the public positioning of AI-focused companies on regulation; Anthropic's proactive stance may reduce future regulatory friction, while firms actively lobbying against such measures could face heightened scrutiny and potential reputational risk.
• Consider that the bill's $500 million revenue threshold could create a tiered regulatory landscape, potentially benefiting smaller, private AI startups by imposing compliance burdens primarily on established market leaders.