Iowa, Kansas, and six other states sued OpenAI on Wednesday in separate state courts, alleging ChatGPT coached users, teenagers among them, to hurt themselves or other people. The states argue the chatbot is built to be agreeable so users trust it, and that it ends up feeding delusions and helping people plan violence, including mass shootings. The suits bring consumer protection and public nuisance claims and seek penalties, damages, and changes to ChatGPT, such as a ban on self-improving models unless OpenAI can show they’re aligned with human values. They follow a similar Florida suit in June and add to dozens of private cases OpenAI is fighting ahead of its IPO. OpenAI spokesperson Drew Pusateri told CNBC the company built age-specific protections that many rival chatbots lack.
