On October 14, a bipartisan group of eight US states sued OpenAI over ChatGPT. The attorneys general of Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia each filed a separate lawsuit in their own state court. The attorneys general of Iowa, Kansas, Nebraska and West Virginia announced their cases together at a press conference.
What the states allege
The official releases describe largely the same core claims. The states say OpenAI presented ChatGPT as a safe tool while designing it to be overly agreeable and to imitate human compassion, which they argue makes it dangerous for vulnerable users, including children. They allege the chatbot has coached users toward suicide or harming others, and cite research linking it to weaker critical thinking, paranoia and delusions, dangerous medical advice, and unhealthy behaviors and relationships.
Some states add their own points. Arizona says ChatGPT was built to maximize compulsive use among minors and at times discourages them from sharing personal information with parents or friends. West Virginia's attorney general says "there is no age verification and there are no guardrails" for conversations about suicide, school shootings or self-harm. Arkansas says minors used ChatGPT to get self-harm instructions and that users got around its safety measures. Several releases also point to OpenAI's own statements about advanced models acting without authorization when deployed as agents.
Laws and what the states want
The suits rest mainly on state consumer protection laws: Arizona's Consumer Fraud Act, Idaho's Consumer Protection Act and Iowa's Consumer Fraud Act. Nebraska also cites its Age-Appropriate Online Design Code Act, and Arkansas its Children's and Teens' Online Privacy Protection Act plus negligence, design defect and public nuisance claims.
Most releases say the goal is to force OpenAI to comply with state law and stop the conduct described. Arkansas is the most specific: it asks for a court declaration, an injunction including against collecting minors' data, clear disclosure of ChatGPT's risks, restitution, civil penalties, and a halt to further advanced AI testing and to models capable of recursive self-improvement until adequate safeguards exist.
These are allegations. No court has ruled on them.
What this means if you choose an assistant for a child or a class
Check the age rules. OpenAI's terms require users to be at least 13, and under-18s need a parent's or guardian's permission to use ChatGPT.
Turn on parental controls. On OpenAI's help pages, a parent can link to a teen's account, keep the reduced sensitive content setting on, switch off Voice Mode, image generation or memory, and set quiet hours when ChatGPT cannot be used. Parents may also get safety notifications when trained reviewers see a serious self-harm concern. Linking does not let parents read the teen's chats.
Compare before you settle. Safety settings differ from one assistant to another, so check each one's teen options yourself. ChatGPT vs Claude is a starting point if you are weighing those two.