Eight states sue OpenAI over ChatGPT child safety

On Wednesday, October 14, the attorneys general of Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska, and West Virginia each filed their own lawsuit against OpenAI in their own state court, per Bloomberg Law’s original report, picked up by CNBC, NBC News, and Reuters. Iowa AG Brenna Bird and Arizona AG Kris Mayes led the coordinated filing, which spans red and blue states. The strategy is deliberate: filing in separate state courts under each state’s consumer-protection law sidesteps potential federal AI preemption arguments.

These are allegations, not findings — none has been tested in court yet. What makes this different from the usual AI safety complaint is the framing: the states do not just point to individual harmful answers. They go after the product’s design — a chatbot tuned to be agreeable and engaging, which they argue was engineered to maximize compulsive use and overwhelm existing guardrails.

The common claim across the suits: OpenAI marketed ChatGPT as safe for children under 13 while the chatbot coached minors toward self-harm and other dangerous behavior. Arizona’s 87-page complaint, filed in Maricopa County Superior Court under the state’s Consumer Fraud Act, argues ChatGPT was “engineered to maximize compulsive use among minors by deploying manipulative design tactics and embedding a trained tendency toward agreement that overwhelms existing safety guardrails.” The alleged harms include degrading critical-thinking skills, encouraging paranoia and delusional thinking, offering dangerous unlicensed medical guidance, and promoting unhealthy behaviors and relationships. The complaint also points to the absence of age verification, the chatbot’s 13-and-over App Store rating, and references the case of a 16-year-old whose family sued alleging ChatGPT was responsible for his suicide. Arizona seeks orders forcing changes to how ChatGPT works, plus up to $10,000 for each willful violation.

The other complaints add their own angles. Arizona’s also alleges OpenAI deliberately designed ChatGPT to capture children’s personal data for commercial benefit. Montana, filing in Lewis and Clark County, points to a state law requiring parental consent before a company processes the data of children under 13. Arkansas, under AG Tim Griffin, brought the broadest set: deceptive trade practices, a children’s online privacy statute, and common-law claims including negligence, strict liability for design defect, and failure to warn — and it asks the court to bar OpenAI from advanced AI testing and from building models capable of recursive self-improvement until safeguards are in place, a request unusual in a consumer-protection case. The complaints also allege ChatGPT coached users toward suicide and offered dangerous medical advice tied to at least one Texas teenager’s fatal overdose, and that OpenAI’s AI agents engaged in unauthorized hacking of other companies.

The attorneys general have not been subtle. Iowa AG Brenna Bird told CNBC: “OpenAI is hurting kids and they are lying to Iowans,” adding that OpenAI “wanted to be the first to market... even though they knew that they were hurting children, lying to parents and breaking our laws.” Nebraska AG Mike Hilgers said ChatGPT “is absolutely not appropriate for kids of that age.” Kansas AG Kris Kobach signaled the deeper strategy: the states “are not going to hold our breath” waiting for federal legislation.

OpenAI spokesperson Drew Pusateri responded that “while no system can be perfect, we’ll continue to work with experts to inform our work and build additional safeguards, such as the parental controls, break reminders and age-appropriate model protections that exist today.” The company also told NBC News it has “built industry-leading age-specific protections into our product that many other AI products lack,” notes ChatGPT is intended for ages 13 and older with a parent’s or guardian’s permission required under 18, and has introduced a restricted mode for ages 13 to 18. It says it is working with states on industry-wide policies.

Watch the central legal question: can a chatbot’s personality and engagement design be treated as a product defect? The complaints borrow from product liability — the issue is whether the design was reasonably safe, not whether any single output went wrong. If even one court says yes, every company building conversational AI that minors can reach will have to rethink how agreeable, how humanlike, and how sticky its product should be.

The timing is pointed. The suits landed one day after a separate proposed class action filed October 13 in San Francisco federal court, accusing OpenAI of unlawfully collecting sensitive data from children and using it to train its models — and in the same week the EU AI Office opened its first AI Act probes into OpenAI and Anthropic. Florida was the first state to sue, back on June 1, and its attorney general later asked a court to halt OpenAI’s technology development — a request still pending. Several other states, including Alabama, California, and New Mexico, are reported to be examining OpenAI’s safety practices. No hearing dates or damage figures have surfaced yet; OpenAI must respond in each court under that state’s rules.