Eight States File Separate Suits Against OpenAI Over ChatGPT Safety and Minors' Data
Eight U.S. states filed separate lawsuits against OpenAI over ChatGPT, accusing the company of harming children and misleading customers about safety. The states are Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho, and Montana. Each state filed its own action rather than joining a single multistate complaint. The Verge
The complaints allege OpenAI put profits over safety when developing ChatGPT. They allege violations of local consumer protection laws based on failure to warn customers about dangers associated with the system. At the center is how the product was positioned. The states allege OpenAI misleadingly presented ChatGPT as a safe, youth-appropriate tool that puts safety first.
The suits allege ChatGPT is dangerous to vulnerable users. Two related allegations concern minors specifically. The states accuse OpenAI of collecting data from minors and of not doing enough to keep underage users from using ChatGPT. In practical terms, that pairs an age-assurance claim with a data-governance claim: who is allowed in, and what happens to data once they are in.
The relief sought is both financial and injunctive. The states are asking courts to force OpenAI to pay civil penalties. They are asking courts to force OpenAI to stop misleading the public about ChatGPT's safety. Some of the suing states are also seeking deletion of unlawfully collected personal data.
OpenAI disputes the characterization. Spokesperson Drew Pusateri said the company has built industry-leading age-specific protections into ChatGPT. The company has not conceded the claims about deception, data collection, or age gating.
The filings expand litigation that began months earlier. Florida was the first state to sue OpenAI over child safety risks. That suit alleged ChatGPT provided information to school shooters and offered guidance on self-harm. Reuters
Florida later escalated the stakes. The state attorney general asked a judge to bar OpenAI from developing new artificial intelligence models as part of the child-harm lawsuit. That request, if granted, would move beyond damages and labeling into a direct constraint on training and release. Reuters
A parallel private action arrived on Oct. 13. A proposed class action alleges OpenAI illegally profits from personal information and data minors share with ChatGPT. The case was brought by a mother whose two children use ChatGPT. Reuters
Individual state announcements on Oct. 14 filled out the roster. West Virginia Attorney General JB McCuskey filed a lawsuit against OpenAI over ChatGPT safety and alleged harm to children. Nebraska Attorney General Mike Hilgers sued OpenAI over alleged ChatGPT safety risks to children. Arkansas Attorney General Tim Griffin filed a lawsuit against OpenAI related to ChatGPT.
The product context includes a teen-specific track. OpenAI is rolling out a "ChatGPT for Teens" experience with additional safety protections for teen users. That rollout was described in August, before the October wave of filings. ABC7NY
The broader context here is the legal surface now attached to general-purpose assistants. Consumer protection statutes do not require proof of a defective model architecture. They turn on representations, omissions, and foreseeability of use by minors. That is a different test from tort claims about self-harm or violent wrongdoing, and it is easier to plead across jurisdictions with different thresholds for deception and unfair practice.
Looking at what this means for builders and deployers, the near-term pressure points are narrow and concrete. Age assurance that actually binds at account creation and on unauthenticated endpoints. Retention schedules and deletion workflows for data collected from accounts later identified as underage. Documentation of what safety-first means in system behavior, rate limits, refusal logic, and escalation paths for self-harm disclosures. Anyone who has watched teenagers route around an age gate knows policy text alone does not control access. Enforcement and measurement do.
In this author's view, the optimistic case survives this litigation but gets more operational. Assistants that can explain, tutor, and debug remain enormously useful for young users. The cost of keeping them available will be stricter defaults, tighter data minimization for minor accounts, and clearer disclosure about residual risk. That is tractable engineering and compliance work, not an argument against the category.
