Woman sues ChatGPT, alleges AI helped ex-boyfriend stalk her despite red flags

AI has become an everyday part of our lives, be it for solving a math equation or making a trip schedule.


But before this technology overtakes every aspect of life, skepticism about privacy policy and data safety has persisted.


Recently, a shocking lawsuit against OpenAI has once again pulled questions about AI safety into the spotlight, raising alarms about how chatbots handle users spiraling into harm.


A California woman has claimed that the platform failed to act on clear red flags, letting a troubled ex's obsession fester with AI fuel.

AI aids ex- boyfriend stalk ex- girlfriend

A California-based woman named Jane Doe, has sued OpenAI in San Francisco court. She claims ChatGPT enhanced her ex-boyfriend's harmful delusions.

According to a Techcrunch report, the 53-year-old Silicon Valley businessman became fixated after heavy use of GPT-4o. He believed he had found a cure for sleep apnea and that secret forces were watching him, according to the lawsuit.


When Doe urged him to get mental health support, he turned even more to ChatGPT. The suit says the AI sided with him rather than urging caution.

Delusions and harassment

ChatGPT reportedly called him “a level 10 in sanity,” sided with his breakup view, and labeled Doe manipulative. He then created fake psychological reports via AI and sent them to her family, friends, and boss to harass her.

The lawsuit argues this AI output justified his stalking, turning digital chats into real torment.

OpenAI safety lapses

OpenAI's automated system flagged the man's account for “Mass Casualty Weapons” activity and temporarily suspended it.

The next day, a human reviewer reactivated it, even though chat titles included “Violence list expansion” that listed specific targets.


Doe submitted a formal abuse report in November, which OpenAI acknowledged but did not act on. The lawsuit alleges OpenAI overlooked at least three clear warnings about the danger.

The man was arrested later

In January 2026, police arrested him on four felonies, Deemed unfit for trial, and entered a mental health facility.


OpenAI paused the account after the suit but denied broader requests like chat log preservation.

What does Doe say?

Doe seeks punitive damages and a court order for OpenAI to save all user chats and alert her to access tries.