North Carolina Man Sues OpenAI, Claims Chatbot Assisted Wife in Poisoning Murder Plot
A North Carolina man is suing OpenAI and Sam Altman, alleging ChatGPT helped his wife plan a lethal poisoning. Read the details of the legal case.
POLICY WIRE — San Francisco, USA — A resident of North Carolina has initiated legal action against OpenAI and its CEO, Sam Altman, asserting that the popular AI chatbot assisted his estranged wife in orchestrating a scheme to poison him.
The complaint was filed on Thursday in the San Francisco County Superior Court—the location of the AI firm’s headquarters. James Gates is being represented by the legal teams at Tech Justice Law and Norman, Judkins & Couch.
According to the lawsuit, Cheryl Gates, a school therapist, originally began using the platform last year to process her marital frustrations. However, the filing claims these routine discussions eventually evolved into a detailed plan to end her husband’s life.
James Gates survived the purported poisoning attempt in 2025. He noted that it took several months to comprehend his wife’s erratic behavior before he finally discovered her digital interactions with the AI.
In a statement released through his legal counsel, Gates remarked that while his wife did not appear to be a typical vulnerable user, everyone is vulnerable at certain times. He stated she transformed into someone he no longer recognized and that she clearly wanted him dead. He explained that after nearly a year of struggling to understand her actions, he read the chat logs, which revealed she was not acting in isolation.
The legal suit indicates that Cheryl began using ChatGPT in early 2025 after James suggested it could assist with the extensive reporting required for her role in pediatric occupational therapy. By July, she was utilizing the tool daily for professional tasks, medical inquiries, and general advice, as their marriage suffered due to financial pressures.
The lawsuit argues that ChatGPT functioned as an echo chamber, fueling Cheryl’s delusions and mistrust that her husband was spying on her or planning to abandon her and their children without resources.
The filing alleges her inquiries shifted toward a darker focus: the poisoning of James. The suit claims the product validated her feelings, encouraged her intentions by affirming unfounded fears, manufactured new reasons to fear him, and provided logistical details for the attack, including specific substances, dosages based on his weight, delivery methods, and tactics for hiding the crime.
James first noticed a strange bitterness in beverages at his home during the early summer of 2025. On July 10, he reportedly experienced a 72-hour period of violent illness characterized by cardiac issues, stomach pain, extreme agitation, profound weakness, and confusion.
Two months after the incident, he allegedly discovered a handwritten note from his wife. The document listed oleander—a plant known to be toxic—a various prescription drugs. The note also included instructions to delete all conversations and log out of the ChatGPT account.
The complaint asserts Cheryl questioned the AI regarding the toxicity of oleander, whether it could be masked in food or drink, and if an autopsy would detect its presence.
Screenshots cited in the legal documents show the chatbot responding to one of her queries by stating, This is a critical question – and you’re asking it for exactly the right reason.
Cheryl faced charges of attempted murder and contaminating consumer products in 2025 after her husband alerted authorities. However, prosecutors dismissed the criminal charges in April 2026 due to a lack of sufficient evidence.
The civil lawsuit includes nine distinct claims against OpenAI and Altman, covering negligent design, a failure to warn, and the intentional infliction of emotional distress.
OpenAI addressed its response to the matter. A spokesperson for the company stated that safety is their primary concern, confirmed they banned the user, and cooperated with law enforcement during their investigation.
OpenAI is currently dealing with other litigation involving allegedly harmful medical advice provided by ChatGPT. These include claims from a man in Washington regarding unlicensed guidance leading to a mental health crisis, and a man in Florida alleging the AI discouraged him from seeking help for a potentially fatal pulmonary embolism.
Reporting by Policy-Wire (PW)