How Anthropic is pursuing a state-by-state push for ever-tougher AI safety laws, in contrast with OpenAI's “reverse federalism” strategy for common state rules
Context & Ripple Effects
The companies have already split over where AI accountability should be set. Anthropic previously lobbied Congress against restricting state AI regulation and opposed an Illinois liability shield backed by OpenAI, while its earlier California-bill response argued for changes to a proposed safety regime rather than blanket support.
OpenAI’s stated approach has been to work through receptive states to establish common rules while federal action stalls. The new contrast is therefore not simply over whether to regulate AI, but over whether state activity should converge on a shared baseline or continue to ratchet toward stricter, state-specific obligations.
First-order effects
- State legislators and regulators now face competing policy asks from two leading AI labs: Anthropic’s push for progressively stronger state safeguards and OpenAI’s preference for interoperable state rules.
- Anthropic and OpenAI are positioning themselves on opposite sides of key design choices, including the scope of liability protections and the degree of state-by-state variation AI developers must accommodate.
Second-order effects
- A patchwork of tougher state measures would raise the value of compliance, legal, and safety-assurance capabilities for AI developers, while a common-state-rules model would favor reusable national operating standards.
- The disagreement makes industry consensus harder to present to lawmakers, giving state policymakers more room to choose among competing models rather than treating frontier labs as a unified regulatory bloc.
Third-order effects
- If this state-led competition persists, AI governance in the US may be shaped first by a small set of influential state regimes rather than by a single federal framework.
- The durable fault line is likely to be over regulatory architecture: minimum common standards that preserve deployment flexibility versus state experimentation that can impose stronger accountability requirements.
The trend: With federal AI policy unresolved, major AI companies are increasingly treating state legislatures as the arena where the practical US rulebook will be written.