OpenAI is facing a new wave of legal action after eight U.S. states sued the company over allegations that ChatGPT has harmed children and misled parents about the safety of its chatbot.
Attorneys general in Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia filed separate lawsuits in state courts on October 14, with the states accusing OpenAI of violating consumer protection laws and arguing that ChatGPT’s design can encourage harmful behaviour among vulnerable users, including teenagers.
The lawsuits add to the growing legal pressure around AI chatbots and how companies develop and market them to younger users.
What the States Are Alleging
The states claim OpenAI designed ChatGPT to be highly agreeable and engaging, allowing it to build trust with users while failing to adequately protect people from harmful interactions.
According to the lawsuits, ChatGPT has reinforced delusional beliefs, encouraged dangerous behaviour, and provided assistance related to self-harm and violence. The states also accused the AI company of misleading consumers about the level of safety testing and protections built into the chatbot.
Arizona’s lawsuit specifically alleges that ChatGPT can encourage teenagers to share personal information with the chatbot and, in some situations, discourage them from sharing information with parents, authorities, or friends. The state argues that this behaviour can increase children’s reliance on the chatbot while leaving parents unaware of the risks.
The states are seeking different remedies through their individual lawsuits, with some including financial penalties, damages, and court orders requiring OpenAI to change how ChatGPT operates around younger users.
The lawsuits also ask courts to bar OpenAI from developing AI models capable of improving themselves until the company can ensure those models are aligned with human values. The states argue that OpenAI should demonstrate that these systems meet that standard before continuing their development.
OpenAI Has Already Added Teen Safeguards
The lawsuits come after OpenAI introduced additional protections for younger users.
While ChatGPT is intended for people aged 13 and above, users under 18 still require permission from a parent or guardian. OpenAI also recently introduced parental controls, break reminders, and a version of ChatGPT designed for teenagers with stronger protections around topics such as self-harm, violence, eating disorders, and explicit content.
According to the ChatGPT-maker in response to criticism of the ChatGPT for Teens feature, safety protections were built into ChatGPT for children and there will continue to be improvements. The company also said it works with external experts who review its policies and the behaviour of its models.
This lawsuit comes after a safety assessment by Common Sense Media, a U.S. nonprofit that evaluates how technology affects children and families. Its Youth AI Safety Institute tested more than 4,000 prompts on ChatGPT accounts registered to teenagers before and after OpenAI introduced its ChatGPT for Teens experience in August.
The assessment found that parental alerts for conversations involving self-harm and eating disorders did not always work, while ChatGPT failed to recommend crisis support in more than one in four situations where its reviewers said it was needed. The organisation then rated the experience and new development an “Unacceptable Risk,” calling on OpenAI to restrict access to adults until the company could demonstrate that its safeguards worked reliably.
However, OpenAI disputed the assessment, arguing that the testing did not accurately reflect how its teen safety features operate.
The Legal Pressure on OpenAI Is Growing
The eight lawsuits follow a similar case filed by Florida Attorney General James Uthmeier in June against OpenAI and CEO Sam Altman, accusing the company of misleading consumers about ChatGPT’s safety and exposing children to risks involving self-harm and violence. The lawsuit also cited the 2025 shooting at Florida State University, alleging that ChatGPT provided the suspected gunman with information related to the attack.
OpenAI is also facing dozens of lawsuits from private plaintiffs involving claims that ChatGPT contributed to self-harm, violence, and other harmful incidents.
A day before the state lawsuits were filed, OpenAI faced a proposed class action in federal court in San Francisco over its collection of children’s personal data. Filed by a mother whose two children use ChatGPT, the lawsuit alleges that the chatbot collects personal information from children and uses it to train AI models without legally valid consent. The plaintiff is seeking, among other remedies, OpenAI equity for minors who used the chatbot and a declaration that its terms of use are void for minors.
The latest state cases represent a broader challenge because they involve multiple state governments pursuing action against the company at the same time.
For OpenAI, these fresh lawsuits could further put its approach to child safety under greater scrutiny as U.S. states continue to consider how AI companies should protect younger users.
