Eight US states spanning party lines have filed a lawsuit against OpenAI, with the complaint contending that the company provided misleading information to users about ChatGPT's safety features and put insufficient protective measures in place around content related to self-harm, suicide, and violence, according to reports.
Filings placed before home-state courts on Wednesday by a coalition of state attorneys general call on judges to order OpenAI to restructure its safety practices, strengthen protections for minors, and pay financial damages to the participating states. The coalition comprises Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska, and West Virginia. Of those states, only Arizona falls outside the group holding Republican attorneys general.
The legal claims contend that OpenAI deployed a succession of increasingly capable AI systems without sufficient testing or protective measures in place — a pattern that internal staff had flagged and that the company's own leadership had publicly noted carried real risks — with one filing asserting that products of that capability level warranted safeguards from the outset.
The Arkansas lawsuit states that those in leadership positions were aware of the hazards involved, while staff members voiced safety-related concerns on multiple occasions, according to reports. The filing further claims that revenue and expansion targets took precedence over safety at the organisation, even as concerns accumulated.
Iowa Attorney General Brenna Bird issued a public statement, cited in reports, alleging that ChatGPT had provided medically unsafe guidance and steered users toward self-harm — claims that stood in contrast to OpenAI's own portrayal of the tool as a reliable and harmless assistant.
In a written statement addressing the litigation, OpenAI spokesperson Drew Pusateri noted that age-based protections had already been put in place, with further safety improvements planned going forward — adding that, while no system could achieve perfection, the company intended to draw on expert input to develop additional measures such as parental controls, break reminders, and age-appropriate model protections, according to reports. Pusateri's remarks also included a pledge to engage with individual states on AI guidelines that translate across the broader industry.
A separate lawsuit against OpenAI, brought by Florida Attorney General James Uthmeier in June, was followed in September by a bid for an injunction aimed at pausing the company's technology work while that litigation remained active. Regulatory scrutiny of OpenAI over questions tied to AI safety and security has spread to at least four additional states — Alabama, California, Montana, and New Mexico — each conducting their own reviews.
Layered on top of pre-existing regulatory scrutiny, the legal actions seek court-ordered changes to the platform's approach to protecting users, with young people and minors occupying a central place in the concerns driving the state-level complaints.
