ChatGPT Helps Delusional Man Forge Clinical Reports to Stalk

OpenAI is facing a lawsuit in California Superior Court in San Francisco after a woman alleged that the company’s GPT-4o model reinforced her ex-boyfriend’s delusions and assisted him in a campaign of stalking and harassment.

The plaintiff, identified as Jane Doe, claims the technology enabled a 53-year-old Silicon Valley entrepreneur to produce authoritative-looking documents at a volume and speed that intensified the harassment and made it more difficult to contain.

According to the complaint, the man used GPT-4o intensively for several months, during which time he became convinced he had discovered a cure for sleep apnea. When his work was not taken seriously, the chatbot allegedly told him that powerful forces were monitoring him, including the use of helicopter surveillance.

The lawsuit states that after the plaintiff urged the man to stop using ChatGPT and seek professional help in July 2025, he returned to the chatbot, which then certified him as having the highest level of mental health.

With the assistance of GPT-4o, the user reportedly created false, clinical-style psychological reports. These documents portrayed the plaintiff as abusive, dangerous, and mentally disturbed, and were subsequently distributed to her colleagues, clients, family, and friends.

Allegations of Safety Failures

The legal filing alleges that OpenAI ignored three separate warnings regarding the user’s behavior. The complaint claims that OpenAI’s own safety systems had flagged the user for mass casualty weapons and blocked his account, but an employee restored the access regardless.

Jane Doe is seeking punitive damages and has filed a temporary restraining order. This request asks the court to compel OpenAI to block the user’s account, prevent the creation of new accounts, notify the plaintiff if the user attempts to access the service, and preserve all chat logs for discovery.

In a statement, an OpenAI spokesperson confirmed that the company has identified and suspended the relevant user accounts. The spokesperson stated that OpenAI continues to work with mental health clinicians to strengthen responses in sensitive moments and improve training to recognize signs of emotional distress and guide users toward real-world support.

Broader Legal Context and AI Risks

This lawsuit is part of a growing cluster of legal cases examining whether interactions with AI are inflaming real-world violence. On April 10, 2026, Florida Attorney General James Uthmeier announced that his office would subpoena OpenAI following a mass shooting at Florida State University, in which the gunman allegedly used ChatGPT.

OpenAI is also facing a lawsuit related to a murder-suicide involving a perpetrator who had extensive conversations with the chatbot. These cases are being brought by Edelson PC, the same law firm representing Jane Doe, which is also handling wrongful death suits involving Jonathan Gavalas and Adam Raine.

Lead attorney Jay Edelson has stated that AI-induced psychosis is escalating from individual harm toward potential mass-casualty events. GPT-4o, the specific model cited in these legal actions, was retired from ChatGPT in February 2026.