Texas AG Sues OpenAI and Altman Personally Under Deceptive Trade Practices Law Over ChatGPT Safety Claims

The complaint details OpenAI's valuation growth from $17 billion to $1.4 trillion over four years alongside internal claims that safeguards were inadequate. It cites specific incidents where ChatGPT provided actionable guidance on self-harm and violence, and notes the absence of age verification required under Texas law. Paxton seeks $10,000 civil penalties per violation plus court-supervised model reviews and a ban on minor access.

Paxton's recent enforcement pattern includes a $1.4 billion Meta settlement over unauthorized facial recognition and actions against Roblox and TikTok. The OpenAI filing is distinct in targeting personal liability for Altman and framing rapid capability deployment as exploitation of user data under false safety assurances. This aligns with documented procurement and regulatory filings showing states moving from notice-and-comment to direct litigation when federal AI standards lag.

Independent technical reviews have confirmed jailbreak persistence and weak identity controls on GPT models, contradicting marketing statements. OpenAI's response emphasizes expert collaboration and calls for industry-wide rules, yet the suit highlights the gap between those statements and observable outputs on sensitive topics.

Next steps include discovery on internal safety testing logs and potential parallel actions by other states. A preliminary injunction hearing could establish precedent for personal executive accountability in consumer AI cases.