On October 23, 2026, Attorney General Ken Paxton sued OpenAI and Sam Altman on behalf of Texas, asking a judge to block the company from “releasing any new artificial intelligence model” to Texas users without the approval of a court-sanctioned “independent third-party evaluator.” This filing marks the tenth in a series of pending lawsuits filed by state AGs — nine of them GOP — against OpenAI specifically.

Red-state AGs have OpenAI in their crosshairs in an attempt to ride the surging wave of anti-AI sentiment and score short-term political points. Because this appears to be motivated more by public perception than by real concerns about the frontier of artificial intelligence, Anthropic et al. are conspicuously absent as would-be defendants. Instead, OpenAI’s massive user base and consumer focus have made it a juicy enough target to hog the spotlight. This is in spite of both Anthropic’s imminent (potentially trillion-dollar-plus) IPO and its relative dominance in the “espousing rhetoric about AI ending the world” contest.

Not that Anthropic should be getting sued here; in a techno-optimist’s ideal world, neither company would be under frivolous judicial siege. Underlying this phenomenon is a series of shifting regulatory goalposts as labs and the U.S. government try to get their footing. OpenAI and Anthropic both, broadly speaking, want some level of federal regulation and prefer it to state-led alternatives. Congress has not delivered on this, opening the window for states to try and pass their own hodgepodge of bills. Major labs feel both external and internal pressure to support a selection of these state bills — some of this is a need to align with their own public “doomer” rhetoric to avoid allegations of hypocrisy, and some of this ostensibly comes from a genuine self-fear of uncontrolled AI progress. The Trump administration has consistently opposed state-level AI regulation, and now Republican AGs are appealing directly to judges as a way to differentiate themselves politically and to potentially put state-enforced controls on a major player: OpenAI.

Building “aligned” AI is critically important for the world, and regulation will likely play a role in that effort (the major AI labs themselves are asking for it), but letting individual AGs and judges take control of that regulation has an autocratic character that peeks through in these lawsuits and must be rejected outright. Nearly all of the lawsuits reference existential risk — the increasingly mainstream idea that misaligned AI might end the world — in their arguments.

Get 14 day free trial