A man in North Carolina, United States, has filed a lawsuit against OpenAI, accusing the company’s artificial intelligence chatbot, ChatGPT, of helping his estranged wife plan an alleged attempt to poison him.
James Gates claims that his wife, Cheryl Gates, used ChatGPT over several months to obtain information about dangerous chemicals, potentially lethal drug combinations and ways of acquiring poisonous substances without attracting suspicion.
According to the lawsuit, Gates began experiencing severe stomach pains, heart-related complications, confusion and extreme anxiety in July 2025.
He later allegedly discovered a handwritten note believed to belong to Cheryl, containing a list of substances, instructions on their quantities and a reminder to delete conversations held on ChatGPT.
The discovery reportedly prompted Gates to contact police, who subsequently recovered several items considered suspicious during their investigations, including syringes, medicine droppers and various medications.
Gates is now accusing OpenAI of negligence, arguing that the responses generated by ChatGPT contributed significantly to the alleged scheme to kill him.
He is seeking unspecified financial compensation from the company.
However, the case has taken a different turn following the decision to drop attempted murder charges against Cheryl in April 2026.
Prosecutors reportedly determined that there was insufficient evidence to establish that an attempt to poison Gates had actually taken place.
Cheryl currently faces no criminal charges in connection with the allegations.
The lawsuit therefore raises fresh questions about the responsibility of artificial intelligence companies over how their platforms are used, particularly when users allegedly seek information that could facilitate harm to others.
