ChatGPT has been sued over allegations that it helped a North Carolina woman poison her husband.
A man in the US state of North Carolina has sued OpenAI and its chief executive, Sam Altman, alleging that ChatGPT helped his estranged wife plan to poison him and provided advice on concealing the alleged crime.
James Gates filed the lawsuit in San Francisco County Superior Court in October 2026, accusing the company of failing to intervene in conversations between the chatbot and his wife, Cheryl Gates.
According to the lawsuit, Cheryl initially asked ChatGPT questions about poisoning, claiming she feared her husband might try to kill her.
The complaint alleges that her questions later became more specific, with at least 25 exchanges discussing poisoning methods, dangerous medication combinations and ways to avoid detection.
James claims the chatbot provided information about concealing harmful substances in drinks, including Celsius energy drinks that he regularly consumed.
The lawsuit further alleges that ChatGPT reinforced Cheryl’s suspicions about her husband rather than challenging unsupported claims or directing her towards appropriate assistance.
James alleges that his wife repeatedly attempted to poison him between June and September 2025.
He reportedly became suspicious after noticing an unusual bitter taste in his energy drinks.
During one alleged incident, he became severely ill for about 72 hours, experiencing heart-related symptoms, abdominal pain, weakness and confusion.
According to the complaint, he later discovered a handwritten note listing potentially dangerous substances and medication combinations. It also reportedly included reminders to delete conversations and log out of ChatGPT.
James subsequently contacted police.
The lawsuit says investigators examined a video recorded by his six-year-old son and footage of a suspicious drink. Police also reportedly recovered medication capsules, syringes, other equipment, poisonous seeds and a tracking device hidden beneath his vehicle.
James says he originally introduced his wife to ChatGPT to help her prepare reports for her job as a paediatric occupational therapist.
However, the complaint alleges that she gradually began using the chatbot for personal matters, including medication questions and marital advice.
As the couple experienced financial difficulties and discussed divorce, Cheryl allegedly became increasingly dependent on the chatbot.
James claims ChatGPT reinforced her beliefs that he was abusive, controlling and planning to abandon her and their children financially.
The lawsuit argues that the chatbot acted as an echo chamber, strengthening her suspicions instead of challenging claims that were not supported by evidence.
OpenAI confirmed that it had banned the account involved in the case and provided information to law enforcement.
A company spokesperson told Ars Technica that safety was a priority and that OpenAI had cooperated with investigators.
The lawsuit accuses OpenAI and Sam Altman of failing to introduce adequate safeguards to detect and interrupt conversations involving plans for violence.
It also alleges that changes intended to make ChatGPT less judgmental when responding to users weakened protections against harmful interactions.
