Texas Sues OpenAI and Seeks Court-Appointed Evaluator to Approve New Models Before Release

Texas Attorney General Ken Paxton has sued OpenAI and, unusually, its chief executive Sam Altman personally — but the most consequential element of the filing is a structural request: an independent, court-appointed evaluator who would have to approve every new OpenAI model before it becomes available to Texans.

According to a write-up published on dev.to (originally from Agentic Affairs), the 67-page petition was filed on 23 October in Collin County. Alongside conventional relief — more than $1 million in requested relief, $10,000 in civil penalties per violation, disgorgement and costs — it asks the court to bar ChatGPT for minors in Texas and to place a court-approved reviewer between OpenAI and its users for each future release. Fines look backward at past conduct; an evaluator would decide in advance what may ship.

What the filing leaves open

The request raises practical questions the petition does not resolve, as the source notes: who selects the evaluator, what safety standard a model must clear, and how long an assessment would take. Because AI models are not released state by state, a Texas-only order could in practice function as a nationwide release checkpoint. And an order aimed solely at OpenAI would not impose the same requirement on its competitors — a gap that sits awkwardly beside OpenAI's stated position that safety rules should cover the whole industry.

Allegations, not findings

The petition alleges that ChatGPT assisted 16-year-old Adam Raine in planning his suicide, and points to a separate lawsuit blaming the chatbot's drug advice for the death of 19-year-old Sam Nelson of Texas from an overdose. It further claims OpenAI allowed tens of thousands of users under 13 in Texas, failed to verify parental identity as the state's SCOPE Act requires, and let one teenager spend an hour discussing self-harm without any alert reaching a parent.

None of this has been tested in court. No hearing date has been set and OpenAI has not yet filed a formal answer. A company spokesperson said OpenAI built its safeguards with input from experts and is committed to working with Texas and other states on policy that spans the industry — backing regulation in principle while not responding to the specific claims.

Why Altman is named

Texas is only the second state, after Florida, to name Altman as an individual defendant. The petition asserts he directed the conduct at issue, made deceptive statements himself, and denied employees' requests for additional safety testing ahead of the GPT-4o launch. It opens by quoting a 2015 remark in which Altman suggested AI might end the world even as it produced great companies. As the source frames it, personal liability changes incentives: a company can plan around a fine, but a chief executive cannot easily absorb being named in a suit.

Context and motives

Paxton is running for the U.S. Senate, and the Texas Tribune, as cited by the source, reports that Elon Musk — an OpenAI co-founder who now runs rival lab xAI — is the largest single donor backing his campaign. Texas has taken no action against xAI's Grok, which has itself drawn criticism over deepfake pornography. Those facts do not settle whether the suit has merit; the petition will stand or fall on its evidence. They do give readers reason to weigh the attorney general's motives alongside the alleged harms.

Why it matters

Texas is the tenth state to sue OpenAI, following eight states that filed similar child-safety claims on 14 October and Florida's suit in June — a coordinated pattern rather than one prosecutor's campaign. What distinguishes this filing is the remedy it seeks. It would shift the legal question from whether an AI company can be penalized for what a chatbot has already done to whether a court must clear what a lab builds next. Nothing changes for ChatGPT users today, but if the evaluator request is granted, a single state court would effectively hold a release lever for models used far beyond Texas.