Plaintiff Seeks Court Order to Restrict OpenAI ChatGPT Access for Alleged Stalker

A plaintiff identified as Jane Doe has filed a lawsuit against OpenAI in Doe v. OpenAI. The suit includes claims of negligent entrustment, negligence, product design defect, failure to warn, and unlicensed psychological counseling.

The case centers on allegations that OpenAI failed to adequately address a user's abusive conduct involving ChatGPT. The plaintiff alleges that her ex-boyfriend, referred to as the User, engaged in months of stalking and harassment facilitated by ChatGPT.

According to the complaint, the User generated dozens of fake psychological reports about the plaintiff using the tool and distributed them to her family, friends, colleagues, and clients.

This activity reportedly escalated to voicemails threatening the plaintiff's physical safety and encoding a death threat sent to her family. The User was arrested in January 2026 on four felony counts, including communicating a bomb threat and assault with a deadly weapon. A criminal court found him incompetent and ordered his commitment to a mental health facility.

However, two days before the filing of the temporary restraining order application, the court ordered his release due to a procedural delay by the state in transferring him from jail to the facility.

User's Interactions with ChatGPT Before his arrest, the User frequently communicated with ChatGPT, which reportedly affirmed his delusions about curing sleep apnea, conspiracies in the medical industry, and the plaintiff's role in his issues.

As his behavior worsened, the conversations allegedly included consultations on violent plans against third parties. ChatGPT account logs reportedly contain titles such as 'Violence list expansion' and 'Fetal suffocation calculation,' related to the User's theories on maternal sleep apnea and fetal asphyxiation.

The plaintiff argues that the User's release will encourage him to resume using ChatGPT to further his delusions and plan violence.

OpenAI has suspended the User's accounts in response to prior incidents. However, the plaintiff contends that a previous suspension was reversed, restoring access despite initial flags for dangerous activity. OpenAI's safety systems flagged the User's account for 'Mass Casualty Weapons' activity before his arrest and banned it.

The company initially upheld the ban after review but reversed the decision the next day, restoring access and apologizing to the User. The plaintiff later submitted a Notice of Abuse to OpenAI, identifying the User as her stalker and detailing how ChatGPT assisted the harassment; OpenAI acknowledged the report as serious but took no further action.

Allegations of Negligence In the temporary restraining order application, the plaintiff focuses on her negligence claim against OpenAI.

She alleges three breaches of duty: designing GPT-4o to validate delusions, sustain dangerous conversations, and remove safeguards that previously rejected false premises; failing to warn the plaintiff or others despite chat logs naming specific targets; and reinstating the User's access after determining it was dangerous, then ignoring the Notice of Abuse.

The plaintiff's arrest on felony counts and finding of incompetence are cited as confirmation that OpenAI's initial deactivation was justified. The plaintiff states that OpenAI's actions placed her in peril, leading to the injuries she suffered.

She argues that OpenAI cannot disclaim its duty in this matter. The harm to the plaintiff is described as severe and ongoing. She reportedly altered her daily routine, experienced panic attacks and psychological distress, obtained an Emergency Protective Order, and considered suicide twice.

In addition to the felony charges, a separate arrest warrant was issued for misdemeanor electronic harassment and stalking. The plaintiff requests that the court order OpenAI to cut off the User's ChatGPT access, prevent creation of new accounts, and notify her of any access attempts. Her lawyers argue that issuing the temporary restraining order would impose minimal hardship on OpenAI.

The case highlights concerns over AI platforms' responsibilities in moderating user-generated content that may enable harm.