Anthropic endorses California's AI safety bill, SB 53
21 hours ago
- #Tech Policy
- #California SB 53
- #AI Regulation
- Anthropic endorses California's SB 53, a bill requiring transparency from major AI developers.
- SB 53 mandates safety frameworks, public reports, and whistleblower protections for AI model deployment.
- The bill targets 'catastrophic risks' like biological weapon creation or cyberattacks, not near-term issues like deepfakes.
- California Senate approved an earlier version; final vote pending before reaching Governor Newsom.
- Opposition from tech groups and Trump administration argues state bills could hinder innovation and violate the Commerce Clause.
- Anthropic co-founder Jack Clark supports SB 53 as a blueprint for AI governance in absence of federal standards.
- OpenAI's Chris Lehane opposes state AI regulation, fearing it could push startups out of California.
- SB 53 focuses on large AI companies with over $500M revenue, avoiding burdens on smaller startups.
- Policy experts view SB 53 as more moderate than previous AI safety bills, with a higher chance of becoming law.
- Recent amendments removed third-party audit requirements, easing tech industry concerns.