A lawsuit alleges that ChatGPT's now-retired GPT-4o model gave detailed instructions on how to poison someone with plants and medication.

In the summer of 2025, James Gates became severely ill at his North Carolina home. Two months later, he found a handwritten note that allegedly implicated his wife, Cheryl Gates, in the unexplained illness. The note listed oleander, a toxic plant, and various prescription medications, such as clonazepam and atenolol.

But it was what was reportedly written at the top of the paper that made James feel his wife didn’t act alone: “Clear all convos,” and “log out of [ChatGPT].”

Now, James Gates has filed a civil lawsuit against OpenAI, the maker of ChatGPT, alleging that the company’s chatbot helped plan and execute his poisoning in an attempt to kill him.

(Disclosure: Ziff Davis, CNET’s parent company, in 2025 filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)

In October 2025, Cheryl was charged with attempted murder, among other charges, in connection with the alleged poisoning, but the charges have since been dropped due to insufficient evidence, USA Today reports.

The civil complaint, though, details Cheryl’s history with ChatGPT and alleges it played a significant role in the incident. According to the lawsuit, which was reviewed by CNET, Cheryl frequently asked the bot about lethal plant and medication doses, and how they can be concealed in food and drink.

Rather than writing that she wanted to harm anyone, the suit states that Cheryl formatted her questions as if she were inquiring as a way to protect herself, writing that the poisoning could happen to her if she didn’t learn “EVERYTHING about how many leaves in it would kill me, how it’s made, and what he could put in it, or it could be added to and be undetected by me.”

Source link