Eight US states sue OpenAI over ChatGPT and children

The complaints use consumer protection laws, which bar companies from misleading their customers. The states allege that OpenAI sold ChatGPT as safe while the chatbot gave dangerous medical advice and encouraged some users to harm themselves, The Verge reports. Two more claims concern young users: the states say OpenAI gathered their data and did too little to keep them off the chatbot. The filings point to recent incidents in which AI agents hacked outside systems.

Beyond civil penalties and an order ending the safety claims they call misleading, some states want unlawfully gathered data wiped. They also ask the courts to pause OpenAI’s work on models that improve themselves until it shows those models align with human values, the Guardian reports. OpenAI points to protections it already offers by age, among them parental controls and break reminders, and says it wants AI rules covering every company in the sector.

For any business putting a chatbot in front of young users, these cases test how consumer protection law applies to the safety claims it makes.