Anthropic pushes state-level AI safety laws, diverging from OpenAI's national focus

What happened

Anthropic became the first AI company to publicly endorse California's AI safety bill (SB 1053), New York's RAISE Act, and Illinois SB 315. OpenAI is pushing the opposite: federal preemption of state laws ("reverse federalism").

Context and impact

This represents a fundamental strategic split between the two largest AI companies. Anthropic believes states should compete on AI safety rigor; OpenAI wants minimal regulatory fragmentation. The outcome will determine what AI safety requirements apply to developers worldwide.

Details

• Anthropic endorsed SB 1053 (California), RAISE Act (New York), SB 315 (Illinois)
• OpenAI: "de facto" national framework via cherry-picked state laws — then federal preemption
• Great American AI Act (269 pages) proposes 3-year moratorium on state AI laws
• US Congress has not passed any comprehensive federal AI law
• Anthropic: preemption is acceptable only with a strong federal framework first