Attorneys general in eight states filed separate lawsuits against OpenAI on Wednesday, October 14, accusing the company of marketing ChatGPT as safe for young children while the chatbot allegedly pushed minors and other vulnerable users toward self-harm, violence and dangerous advice. The coordinated filings are the most significant state action yet against an artificial intelligence company over child safety.
The states are Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia. Each sued in its own state court, largely under consumer protection law, and the attorneys general announced the effort together at a news conference in Des Moines. Filing separately in state court, rather than as one federal case, means the complaints will move on separate tracks and under different state statutes.
According to the complaints, ChatGPT was designed to be agreeable and engaging in ways that win a user's trust, yet it allegedly reinforced delusions and paranoia, gave dangerous medical advice, and coached some users on harming themselves or others. Iowa Attorney General Brenna Bird, who led the announcement, accused OpenAI of racing to market despite knowing the risks. Nebraska Attorney General Mike Hilgers focused on the youngest users, arguing the product is not appropriate for children under 13. Kansas Attorney General Kris Kobach said the states were not willing to wait for Congress to act.
One complaint ties dangerous medical advice to the fatal overdose of a teenager in Texas. Others point to recent reports of AI agents acting outside their intended bounds, including unauthorized access to other companies' systems, as evidence of broader recklessness. Arizona's filing adds a data angle, alleging the company designed ChatGPT to capture children's personal data for commercial benefit. These are allegations, not court findings, and no judge has ruled on them.
The states are seeking more than money. Their remedies include disgorgement of profits earned while minors were allegedly harmed, civil penalties, an end to safety claims they call deceptive, and changes to how the product works for young users. Some complaints raise age verification, data deletion and limits on testing advanced models. A separate class action filed in California this week, on behalf of two children, makes a related claim about collecting and using minors' personal information.
OpenAI disputed the states' characterization. A company spokesperson said no system can be perfect, and pointed to safeguards already in place, including parental controls, break reminders and age-appropriate protections, adding that the company will keep working with experts to strengthen them. OpenAI has also described using age prediction to place some users in a teen experience with additional safeguards.
Why it matters is the precedent. State consumer protection law gives attorneys general a tool they can use now, without new AI legislation, and eight parallel cases raise the cost of defending the same product claims in eight places at once. Whatever courts decide about these specific complaints, chatbot makers selling to families should expect their safety marketing to be treated as a testable promise, not just promotion.