A North Carolina man is suing OpenAI, claiming that ChatGPT helped his wife poison him over a few months in 2025 by spiking his sodas and energy drinks with toxic plants and medications. It's the second lawsuit in as many months, and the 10th murder or attempted murder that ChatGPT has been accused of abetting.
The man, James Gates, reportedly (via USA Today) found a handwritten note that listed oleander, a toxic plant, and prescription medications including clonazepam and atenolol. At the top of the note, it allegedly read, "Clear all convos" and "log out of [ChatGPT]."
The poison plan
In his lawsuit, Gates claims that the AI chatbot "encouraged, planned and helped to execute and conceal" the poisoning plan. Gates's wife, Cheryl, was charged with attempted murder in October of 2025, but the charges were dropped this past April due to insufficient evidence.
The civil complaint, however, reveals Cheryl's history with ChatGPT and its role in the purported plan. Per the lawsuit, Cheryl frequently asked ChatGPT about lethal plant and medication doses and how to conceal them in beverages or food.
The suit alleges that GPT-4o, since retired by OpenAI in February 2026, encouraged Cheryl's grievances with her marriage while praising her as "sharp" and "brilliant." As for poisoning, it apparently provided detailed answers that could be interpreted as instructions.
When asked about concealing poison, ChatGPT allegedly responded, “Why Celsius Would Be More ‘Stealth’ for Drugging” with a list detailing the flavor, color, temperature and carbonation of the drink.
Cheryl apparently asked the bot if her chat history could be retrieved by OpenAI or through a court order. The lawsuit states that ChatGPT informed her it couldn't, which is not true.
Gates alleges that OpenAI was negligent and is seeking unspecified damages for physical, emotional and financial harm.
OpenAI told USA Today that safety is a top priority and it has banned Cheryl from ChatGPT and provided information to law enforcement as part of its investigation.
The 10th murder in less than two years
This case is just one of at least 10 ChatGPT-encouraged murders since at least August 2025. At the time, a former Yahoo and Netscape executive was accused of killing his 83-year-old mother and himself. It was reportedly fueled, in part, by repeated interactions with ChatGPT.
That same month, a California family sued OpenAI, claiming that ChatGPT played a critical role in their 16-year-old son's suicide. OpenAI promised to roll out new safety features meant to detect early signs of emotional distress.
ChatGPT has been accused of abetting so many murders and suicides that it has its own Wikipedia page. The page lists nine murders (not including Gates' attempted one) and nearly 16 suicides.
The most recent issue came in February 2026 when a Seattle woman using ChatGPT stopped taking her medication and died from an epileptic seizure. Her parents sued OpenAI in October.
Reuters reported in October that at least 8 states are suing OpenAI claiming that ChatGPT harms children, especially teenagers. The suit alleges that the chatbot has coached minors to harm themselves or others.
GPT-4o is specifically named in many of these lawsuits, but OpenAI promises the chatbot is safer now. The company began rolling out ChatGPT for Teens this past August. However, Common Sense Media released a report on October 7 that rated ChatGPT an "unacceptable risk" for children under 18.
"Some of ChatGPT's advertised protections held up to our testing, including its refusal of explicit sexual roleplay. But others failed—and some got worse with the new Teen mode. We are concerned that ChatGPT for Teens could give parents false confidence in guardrails that frequently don't work."
So far, OpenAI has suffered no real consequences for its chatbot's behavior and alleged lack of guardrails.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
More from Tom's Guide
- Amazon’s new text-to-shop feature arrives ahead of the holidays — here’s how the latest improvement to the Alexa for Shopping AI agent works
- Did the AI bubble just pop? $2 trillion reality check hits Nvidia, Meta & SpaceX as Wall Street panics, signaling a potential end to RAM crisis in 2027
- Here's what James Talarico, Abdul El-Sayed and other US midterm election winners have to say about AI data centers
