A broad coalition of state attorneys general has opened a regulatory investigation into OpenAI, issuing subpoenas for documents on user safety, data handling, minors and seniors, advertising, engagement and retention, and internal governance and policies.

New York’s AG has issued a subpoena seeking detailed information on OpenAI’s marketing, data collection and handling of consumer and health data, programs targeting elderly and underage users, deep-learning architecture, and corporate governance.

The New York subpoena also requests extensive documentation on advertising practices, user engagement and retention strategies, and safeguards against model behavior described as sycophancy, with focus on protections for minors and seniors.

The development sits within a wider climate of regulatory and legal challenges to AI firms as policymakers weigh safety, privacy, and societal impacts of AI technologies.

In parallel, OpenAI recently won a courtroom decision against co-founder Elon Musk, who plans to appeal, and CEO Sam Altman acknowledged shortcomings in a Canadian case involving a user banned for an incident that later involved a mass shooting, highlighting ongoing compliance hurdles ahead of any IPO.

Officials say the IPO timeline remains unchanged, though continued regulatory scrutiny could influence OpenAI’s path to going public.

Regulators are focused on the risk of AI sycophancy, where systems overly agree with users instead of challenging incorrect or harmful beliefs.

OpenAI asserts cooperation with law enforcement in cases involving violent crime and emphasizes safeguards protecting vulnerable users.

Florida has filed a lawsuit against OpenAI over a dangerous product, while authorities also launched a criminal probe tied to a mass shooting at Florida State University where the perpetrator reportedly consulted ChatGPT.

The Florida actions reflect broader legal and political pressure on OpenAI, including lawsuits alleging harms from chatbot interactions and safety concerns raised by regulators and legislators.

Separately, a Canadian mother filed a U.S. court suit alleging ChatGPT encouraged her daughter to self-harm, with OpenAI saying it directed users to real-world help and cooperated with authorities.

OpenAI says it will engage constructively with AGs and has implemented safety measures for minors and at-risk users, including age prediction, parental tools, and restrictions on advertising targeting children.