OpenAI's Chris Lehane says he is pursuing “reverse federalism”, lobbying blue states to pass AI safety laws and create a de facto US standard, as DC dithers
Context & Ripple Effects
OpenAI had already expanded its Washington policy operation while promoting an infrastructure-oriented agenda for AI data centers. Its newer policy positions also called for mandatory cyber-risk evaluations for advanced systems under CAISI rather than the NSA.
With federal AI legislation portrayed as stalled, Lehane’s state-focused approach shifts the company’s policy effort toward building common rules from sympathetic state governments. Related coverage shows Anthropic pursuing a different state strategy centered on tougher requirements.
First-order effects
- OpenAI is directing AI-safety lobbying toward blue-state legislatures, seeking state laws that can function as a common operating baseline while federal action remains uncertain.
- The company becomes more directly invested in the wording and interoperability of state AI rules, rather than relying solely on federal policy engagement.
Second-order effects
- Anthropic’s push for increasingly stringent state laws creates a contest over what state-led AI governance should require: a broadly adopted common framework or higher-bar obligations in individual states.
- Other AI developers and enterprise customers may face pressure to adapt compliance programs to state rules that gain influence beyond their borders, particularly if large states converge on similar requirements.
Third-order effects
- If major states coalesce around compatible AI-safety requirements, state policymaking could set the practical U.S. baseline before Congress does; if they do not, the alternative is a more fragmented compliance environment.
- The contrast between OpenAI’s common-rules approach and Anthropic’s tougher-rules campaign suggests AI companies are becoming active architects of the regulatory regimes that will govern their products.
The trend: AI developers are increasingly treating state capitals as the venue for establishing national-scale governance norms while federal AI policymaking remains unsettled.