Eight US states sue OpenAI over ChatGPT child safety

By Splitfeed EditorialLast updated Oct 14, 2026, 22:40 UTC

On October 14, 2026, eight states, Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho, and Montana, each filed separate lawsuits against OpenAI. The suits allege the company violated consumer protection laws by misleading users about ChatGPT's safety and prioritizing profits over the welfare of children. Arkansas Attorney General Tim Griffin and the Kansas Attorney General each named OpenAI directly for deceptive safety claims. At least one complaint alleges OpenAI used children's personal data, including their fears, hopes, and secrets, to train its models. OpenAI has not issued a public response to the suits.

The coordinated timing of eight simultaneous state-level filings marks a notable escalation. Rather than federal regulatory pressure, these are product liability and consumer protection claims, grounding the legal challenge in existing state law rather than waiting for new AI-specific legislation. As the account London5 noted, the shift from regulatory debate to product liability is significant, because it means OpenAI faces concrete legal exposure in multiple jurisdictions at once, with remedies that could include damages and injunctions, not just rule-making.

The Anti-AI camp has leaned into the specific allegations. RoseMarie quoted Iowa's attorney general as saying OpenAI claims ChatGPT is safe while giving dangerous medical advice, encouraging people to hurt others, and coaching people to commit suicide, and separately that OpenAI is hurting kids and lying to Iowans. Accounts including Privacy Professionals Briefly and EU Data Protection Briefly framed the suits as confirmation of a pattern of misleading safety claims and harmful behavior. The Pro-AI camp disputes the severity framing. Decorative Jord argued that critics hold the contradictory position that these are mere word-guessing machines yet somehow lethal after seconds of exposure. Leshrac suggested ChatGPT is no more dangerous than spending hours on social media. Will Stancil characterized the alarm as closer to psychosis than reasoned concern. The Middle Ground, represented by London5's observation about product liability, leaves the underlying safety questions open while treating the legal mechanism itself as the meaningful development.

The cases will move to discovery, where internal OpenAI documents about safety testing and data practices would become central. Any early motions to dismiss, or a settlement offer from OpenAI, would be the next concrete signals of how seriously the company treats this coordinated challenge.