Scott Wiener on his fight to make Big Tech disclose AI’s dangers

California's proposed AI safety bill, SB 53, is currently awaiting Governor Newsom's decision, aiming to mandate safety reporting for AI companies generating over $500 million in revenue, specifically targeting catastrophic risks such as bioweapons and cyberattacks. This iteration, which also includes whistleblower protections and a state-operated cloud for AI research, has garnered more industry support, including from Anthropic and Meta, by focusing on transparency and self-reporting rather than the liability provisions of its vetoed predecessor, SB 1047. If enacted, SB 53 would establish some of the nation's first state-level mandatory AI safety disclosures, potentially influencing future regulatory approaches despite some industry preference for federal oversight.

Analysis

California's proposed AI safety bill, SB 53, represents a significant, more moderate legislative effort following the veto of its predecessor, SB 1047. Currently awaiting the governor's signature, the bill pivots from imposing broad liability to mandating transparency through safety reporting for AI companies with over $500 million in revenue. This shift has tempered industry opposition, securing an endorsement from Anthropic and qualified support from Meta (META), which described it as a "step in that direction." The legislation specifically targets catastrophic risks like bioweapons and cyberattacks, and includes provisions for employee whistleblower protection and a state-run cloud computing cluster, CalCompute, to support broader AI research. Despite this softer approach, some industry players like OpenAI maintain a preference for a single federal regulatory standard, citing risks of a fragmented state-by-state compliance landscape. If enacted, SB 53 would establish one of the nation's first mandatory AI safety reporting frameworks, creating a key regulatory precedent in a major technology hub and highlighting the growing divergence between state-level safety concerns and the federal government's current focus on AI growth and deregulation.

Key Decisions for Investors

• Investors in large-cap AI developers like Alphabet (GOOGL) and Meta (META) should closely monitor Governor Newsom's decision on SB 53, as its passage would introduce new compliance costs and operational reporting requirements in a key market.
• The bill's focus on firms with over $500 million in revenue could create a competitive advantage for smaller, exempt AI startups, a dynamic that long-term investors should consider when evaluating the AI sector's competitive landscape.
• Factor in the growing regulatory risk from a potential patchwork of state-level AI laws, as the divergence between California's safety-focused approach and the federal government's growth-oriented stance could increase complexity and legal expenses for companies operating nationwide.