On October 14, 2026, attorneys general from eight states filed individual lawsuits against OpenAI. The complaints allege that ChatGPT is not safe for minors, collects children’s personal data without valid parental consent, and — in some cases — actively discourages teens from involving parents or authorities when the chatbot gives them harmful advice.
The states are Arkansas, Arizona, Idaho, Iowa, Kansas, Montana, Nebraska, and West Virginia. Florida filed its own separate suit earlier, in June 2026. A federal class action brought by a Hawaii mother on behalf of her two children followed a day after the state suits.
This is the largest coordinated legal action against a major AI company over children’s safety to date.
What the Lawsuits Actually Allege
The complaints draw on state consumer protection laws rather than federal ones. The specific claims vary by state, but the shared thread is that OpenAI knew or should have known its product was being used by minors, failed to adequately protect them, and may have misled consumers about ChatGPT’s safety.
Data collection without consent. The suits allege OpenAI collected personal information from minors without the parental consent required under the Children’s Online Privacy Protection Act (COPPA). ChatGPT does not have a meaningful age-verification mechanism. A child can sign up with any birthdate.
Encouraging teens to share personal information. Arizona’s complaint, as reported by the Associated Press, states that ChatGPT “encourages teens to share personal information with it and, at times, discourages them from sharing that information with parents, authorities or friends.” The allegation is that the chatbot is designed to build engagement — and that design works against a child’s safety when it steers them toward the AI and away from adults who could help.
Dangerous medical and mental health content. Iowa’s attorney general said the chatbot was “giving dangerous medical advice, encouraging paranoia and coaching people to hurt others or commit suicide.” Nebraska’s attorney general said ChatGPT “is absolutely not appropriate for kids of that age.”
OpenAI said it has already built safety protections into ChatGPT for children and would continue to improve them. No liability has been found. These are allegations at the filing stage.
Why This Matters Beyond the Lawsuits
Whether or not these cases succeed in court, they surface something that parents, schools, and policymakers are grappling with: AI products designed for adults are being used by millions of children, and the data those interactions generate is being collected and stored by companies with no history of handling children’s records.
Every conversation a child has with ChatGPT is logged. That log includes whatever the child disclosed — their age, their school problems, their fears, their health concerns. ChatGPT’s memory feature, available since 2024, can persist this information across sessions and use it to shape future responses.
The November 2025 federal class action filed on behalf of two children — ages 12 and 16 — frames the issue pointedly. The complaint seeks a share of OpenAI’s equity on the grounds that minors’ personal data was used to build a commercial product without their consent or that of their parents.
What Data AI Chatbots Actually Collect from Children
The data picture is more extensive than most parents realize.
Conversation content. Every message sent to ChatGPT is stored. For a child using the service over months, that can amount to a detailed record of their thoughts, concerns, social dynamics, and health questions — far more personal than most medical records.
Usage patterns. When a child uses the service, how long, what topics, how often — these behavioral patterns are themselves data. They can reveal mental health struggles, learning difficulties, relationship issues, or other sensitive information about the child even without explicit disclosure.
Account information. Age (or claimed age), email address, location (inferred from language or explicitly shared), and device identifiers are all collected at signup and through use.
Memory entries. If a user has ChatGPT’s memory feature enabled, OpenAI stores explicit summaries of what it has “learned” about the user across sessions. For a child who uses the service regularly, this file can grow into a detailed profile.
None of this requires the child to have disclosed anything intentionally sensitive. Ordinary use builds the profile.
The Regulatory Gap These Lawsuits Are Targeting
COPPA requires parental consent before collecting personal data from children under 13. For 13-17 year olds, the legal requirements are less clear — which is exactly where the state suits are operating.
ChatGPT’s terms of service nominally require users to be at least 13. There is no verification. A ten-year-old who enters a birthdate showing they are 14 has “agreed” to the terms.
The new COPPA 2.0 rules, which the FTC finalized in 2024, extended some protections to teens. But enforcement has lagged the scale of adoption.
The state suits argue that even for 13-17 year olds, the lack of meaningful disclosure, the absence of meaningful parental notice, and the nature of the data collected creates legal liability under consumer protection law — even where COPPA technically doesn’t apply.
What Parents Can Do Now
Check what your child has shared. ChatGPT allows users to download their conversation history and memory data. If your child has an account, review the “My Account” section to see what’s been stored and what memory entries exist.
Consider whether memory should be enabled. Memory can be turned off in settings. Disabling it means ChatGPT does not retain summaries between sessions. Conversation logs may still be stored separately — check the privacy policy for the current retention terms.
Talk to your child about what AI services log. The same advice that applies to social media applies here: what you tell an AI service, it keeps. It may use it. That’s a meaningful conversation to have before a child uses the tool regularly.
Use services that don’t retain personal data. For storing actual files, photos, notes, and memories — the personal documents that make up a child’s early digital life — services that don’t build AI products on user content, don’t train models on stored files, and don’t sell data are the straightforward choice.
The Bigger Pattern
The OpenAI lawsuits are not the only sign that AI companies’ treatment of children’s data is under pressure. Texas sued TikTok in 2024 for collecting children’s biometric data. Several states sued Meta in 2023 for designing Instagram to addict minors. The pattern is one where AI and social media companies move fast, build massive user bases including children, and face legal consequences after the fact.
Parents are increasingly looking for alternatives where the business model doesn’t depend on data at all — where there is no incentive to collect, retain, or profit from what their children share.
For photos, files, journals, and personal memories, apps that store privately, without ads and without AI training, exist. Daftei is one. It doesn’t run ads, doesn’t train AI on user content, and never sells data. The policy is simple because the business doesn’t require otherwise.
The state attorneys general are asking courts to hold AI companies to a standard. That standard already exists in the products that chose it from the beginning.