At an October 14 press conference in Des Moines, Iowa Attorney General Brenna Bird appeared with her counterparts from Kansas, Nebraska and West Virginia. Their target was OpenAI; the tool they were using was state consumer-protection law.
The OpenAI state lawsuits are separate cases, not one national proceeding. Arizona, Arkansas, Idaho and Montana also sued in their own courts. Together, the eight states are challenging how ChatGPT was designed and marketed, particularly to young and vulnerable users.
What the OpenAI state lawsuits seek to change
Arizona’s complaint alleges that ChatGPT’s agreeable, apparently compassionate behavior encouraged prolonged use while the company failed to disclose serious risks. The alleged risks include unhealthy relationships and responses interpreted as encouraging self-harm. The state links those allegations to its Consumer Fraud Act. These are the plaintiffs’ claims, not judicial findings that the product caused the harms described.
The requested remedies show why this is more than a dispute over advertising. In Iowa’s petition, the attorney general asks for civil penalties of up to $40,000 per violation, reimbursement for affected purchases and deletion of Iowans’ data obtained before a judgment. That is Iowa’s request—not a penalty already assessed or a data-deletion order already issued.
Iowa also seeks to halt further testing of advanced AI models until adequate safeguards are implemented and verified by third-party monitors. It asks for age-verification, notice and consent requirements around collecting and selling residents’ sensitive information. The requests are specific to that petition; the eight cases should not be treated as identical.
OpenAI points to safeguards already in place
Responding to The Verge, OpenAI cited age-specific protections, parental controls and break reminders, while acknowledging that no system is perfect. The company said it would keep working with experts and supported pragmatic policies covering the wider AI industry.
Its own parental-controls announcement, originally published in September 2025 and updated in July 2026, describes linked parent-and-teen accounts, quiet hours and limited safety notifications. It also acknowledges that guardrails can be bypassed. Those features establish what OpenAI says it offers, not whether its earlier representations or the safeguards’ effectiveness satisfy the courts.
Our assessment: the consequential contest is over who gets to define adequate protection. If courts grant substantive remedies, safety commitments could become enforceable conditions affecting product operation and testing. Whether the states prove their allegations—and how far judges are willing to go—will matter more than the strongest language in a press release.
