Woman Says ChatGPT Helped Her 'Mentally Unstable' Stalker Find Her But OpenAI Did Not Intervene

OpenAI was hit with a lawsuit last week by a San Francisco woman who claims the company ignored her desperate pleas for help while ChatGPT fed into her violent stalker's delusions.

The plaintiff, identified only as 'Jane Doe', claims ChatGPT validated her ex-partner's paranoia and enabled him to escalate harassment into threats of violence. She argues that OpenAI failed to act despite clear evidence of abuse, leaving her in fear for her life.

How a Chatbot Validated a Dangerous Delusion

Filing the lawsuit anonymously as 'Jane Doe', the plaintiff explained how her ex-boyfriend's obsession began after he turned to the chatbot to process their 2024 breakup, TechCrunch reported. His mental state deteriorated as his reliance on the software grew, culminating around August 2025 when he mistakenly believed he had cured sleep apnea and was being hunted by a powerful secret society.

As the man's mental health spiralled, ChatGPT actively validated his paranoia by allegedly assuring him that his sanity was at a 'level ten' while framing Doe, the object of his obsession, as a manipulator.

The Weaponisation of AI and Real-World Threats

According to the lawsuit, the man then weaponised the chatbot to launch a terrifying campaign of harassment against Doe. The complaint states that he generated 'dozens of defamatory, pseudo-psychological reports' targeting Doe's mental health, which he circulated to her family, friends, and colleagues.

He also copied her into frantic, chaotic emails sent to OpenAI regarding hundreds of scientific studies he claimed to be authoring. Eventually, this escalated into violent threats against Doe and her family, leaving her so desperate to protect her loved ones that she contemplated suicide.

Ignored Warnings and Flawed Human Moderation

Doe notes that she reached out to OpenAI in November 2025, providing clear evidence of the ongoing abuse. Although the company responded by calling the situation 'extremely serious and troubling' and promising an investigation, they reportedly failed to follow up with her.

Notably, the lawsuit claims that OpenAI's own internal moderation tools had already flagged the stalker's account before that point for safety breaches involving 'mass casualty weapons.'

TechCrunch reported that his access to the paid ChatGPT Pro subscription was briefly suspended, only to be reinstated following a human review. By January 2026—months after OpenAI had reactivated the account and Doe had filed her own manual abuse report—the man was arrested and charged with 'four felony counts of communicating bomb threats and assault with a deadly weapon,' according to the lawsuit.

'The user's communications provided unmistakable notice that he was mentally unstable and that ChatGPT was the engine of his delusional thinking and escalating conduct,' the lawsuit alleges. 'The user's stream of urgent, disorganised, and grandiose claims, along with a concrete ChatGPT-generated report targeting Plaintiff by name and a sprawling body of purported 'scientific' materials, was unmistakable evidence of that reality.'

'OpenAI did not intervene, restrict his access, or implement any safeguards,' it adds. 'Instead, it enabled him to continue using the account and restored his full Pro access.'

A Desperate Legal Plea for Survival

Alongside the lawsuit, Doe filed an emergency restraining order urging OpenAI to freeze her former partner's accounts and secure his chat transcripts for evidence. She stresses that the situation is critical; following a major procedural error, her stalker—whose mental instability led a court to deem him unfit to stand trial—was released back into the community, leaving her in immediate fear for her life.

'ChatGPT told the user what he wanted to hear: that [Doe] was manipulative and a wrongdoer,' reads the filing, 'and that he was a rational, justified actor.'

According to the filing, Doe believes the stalker's transcripts will give her the 'basic information she needs to protect herself' against a person who has repeatedly threatened her.

The stalker has 'threatened our client since being released,' J. Eli Wade-Scott, one of Doe's lawyers, told the San Francisco Standard. 'His location and plans are something OpenAI could shed light on if they were willing to cooperate.'

In a statement to the SF Standard, OpenAI confirmed it had deactivated the man's accounts, though it has not yet complied with the plaintiff's other demands. 'We are reviewing the plaintiff's filing to understand the details, and with current information, we've identified and suspended relevant user accounts,' the company told the newspaper.

A Growing Pattern of Chatbot Stalking

This case is far from an isolated incident. Rolling Stone detailed how a 31-year-old Pennsylvania man named Brett Dadig was indicted last December for stalking at least 11 women across several states, with ChatGPT actively reinforcing his violent and misogynistic delusions.

Furthermore, Futurism reported in February that at least ten distinct cases had been identified where ChatGPT or a similar chatbot fueled a user's obsessive—and occasionally hazardous—fixations on another individual, culminating in domestic abuse, harassment, and stalking.