A North Carolina man has alleged that ChatGPT helped his estranged wife plan an attempt to poison him.
Gates had suffered a severe bout of illness months before finding the note in September 2025, according to the lawsuit. Cheryl began using ChatGPT regularly in 2025, initially seeking help with her work as a paediatric occupational therapist and routine household queries, according to the complaint. Cheryl Gates was charged with attempted murder and food contamination following a police investigation after her husband reported his suspicions in 2025.
A North Carolina man has sued OpenAI, alleging that its artificial intelligence chatbot ChatGPT helped his estranged wife plan an attempt to poison him during their marital troubles, according to a Bloomberg report. Gates said he became suspicious of his wife after discovering a handwritten note containing a list of substances and dosage instructions, some marked as “lethal”. The note also reportedly carried instructions such as “clear all convos” and “log out of ChatGPT”. However, prosecutors dropped the charges in April, citing insufficient evidence, according to court records. The couple is now separated, and James Gates has temporary custody of their two children, according to the lawsuit. “We’re no different than any family out there,” he said. “We banned this user and have provided information to law enforcement as part of their investigation into this case,” the company said.
James Gates claimed in a lawsuit filed in San Francisco County Superior Court that his wife, Cheryl Gates, repeatedly consulted ChatGPT while their marriage was falling apart and used the chatbot to develop a plan to kill him. The lawsuit alleges that ChatGPT helped Cheryl research poisonous substances, determine potentially lethal quantities and explore ways to conceal a poisoning. It also names OpenAI CEO Sam Altman as a defendant. He described experiencing cardiac symptoms, extreme jitteriness, abdominal pain, severe weakness and confusion. However, as tensions grew between the couple and they began discussing separation or divorce, she allegedly started discussing her marital problems with the chatbot. The lawsuit claims ChatGPT reinforced her suspicions that her husband was spying on her and planning to leave her and their children without financial support. Cheryl also allegedly asked the chatbot about oleander, a poisonous plant, while presenting her queries as concerns about protecting herself from her husband. Screenshots cited in the complaint reportedly show conversations in which she sought information about potentially fatal quantities of oleander, how it could be concealed and whether it would be detected during an autopsy. In one exchange, when she reportedly asked whether oleander would appear in an autopsy report, ChatGPT responded: “This is a critical question — and you’re asking it for exactly the right reason. There are currently no criminal charges against her. “If it could happen to us, it could happen to anyone. OpenAI defended its safety measures in a statement to Bloomberg, saying safety remained a priority. The lawsuit adds to a growing list of legal challenges against OpenAI and other AI chatbot developers over allegations that their products contributed to mental health crises, suicides and violent behaviour.
Gates is seeking unspecified damages, alleging that OpenAI was negligent in designing its product and that ChatGPT encouraged, planned and helped conceal the alleged poisoning attempt, USA Today reported. Speaking to USA today, Gates said he had never imagined that his family life could take such a turn.
OpenAI and other technology companies have faced scrutiny over the safeguards built into their chatbots and their responsibility for harmful content generated during conversations with users. You use AI every day. Now get your AI Quotient. Take the AIQ test.