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
OpenAI's effort comes after nearly a year in which the tech lobby has pressured Congress to block states from passing AI laws …
Context & Ripple Effects
OpenAI had expanded its Washington policy operation and, in a March policy submission, warned that the volume of pending state AI bills could weaken U.S. progress. Its current state-focused approach marks a tactical shift from arguing against fragmented rules to trying to shape a common baseline through selected states.
The subsequent coverage contrasts this approach with Anthropic's push for increasingly stringent state-by-state requirements, making the key policy contest less about whether states act than about what level of safety obligation becomes the model.
First-order effects
- Blue-state legislatures become immediate venues for OpenAI's preferred AI-safety framework, while OpenAI can seek to influence the terms of rules that may affect its own systems.
- The move puts OpenAI at odds with the broader tech-lobby campaign for Congress to prevent state AI laws, at least where state action can produce interoperable rules.
Second-order effects
- Anthropic and other AI developers face pressure to engage more directly in state policymaking, both to defend stricter approaches and to avoid having OpenAI's preferred baseline become the reference point.
- A competition between a common-rule approach and tougher state-specific mandates could determine whether companies can standardize compliance programs or must accommodate divergent requirements.
Third-order effects
- If several large states align around similar requirements, state law could function as a practical national AI-governance baseline even without congressional action.
- If alignment fails, the same state-level competition could entrench the patchwork OpenAI previously warned about, increasing the importance of federal preemption or a federal standard.
The trend: AI companies are shifting from lobbying over whether regulation should exist to lobbying over which jurisdiction and which safety baseline will govern the market.