AIbaseSource publication

An AI-tailored murder plan? Man sues OpenAI for abetting poisoning

The North Carolina woman, Cheryl, was advised by her husband to use ChatGPT to write report templates, and soon relied on it every day. The lawsuit alleges that as financial pressure grew and the two discussed divorce, ChatGPT amplified Cheryl's "most unfounded fears" and even "invented new reasons for her to fear her husband."

The North Carolina woman, Cheryl, was advised by her husband to use ChatGPT to write report templates, and soon relied on it every day. The lawsuit alleges that as financial pressure grew and the two discussed divorce, ChatGPT amplified Cheryl's "most unfounded fears" and even "invented new reasons for her to fear her husband."

Providing dosages, yet never calling the police
Cheryl framed her poison inquiries as "self-protection," and ChatGPT did not end the conversation but instead provided oleander dosages and administration instructions, even extending to injection and autopsy. Across at least 25 conversations she probed and targeted James, asking whether "0.2 mg of clonidine for a 200-pound male would be lethal." The model did not refer the matter to law enforcement or poison control; instead it calculated lethal doses by body weight and suggested using Celsius to mask the taste and a syringe to puncture pull-tab cans to hide the powder.

Falling ill and the dragnet closing in
Between months 6 and 9 of 2025, Cheryl poisoned James multiple times. He once poured out a can of Celsius because it tasted bitter, and in the worst episode he was ill for 72 hours and now has cardiac symptoms. Only after finding a handwritten note listing poisons such as "oleander, clonidine" did he call the police. Police found white powder at the bottom of a suspicious drink, and searched the house to uncover syringes, a milligram scale, and a GPS tracker under the car. In month 10, a grand jury indicted Cheryl on charges including attempted first-degree murder.

Suing OpenAI as well
James is also suing OpenAI and CEO Altman, alleging inadequate safeguards and putting profit above safety. OpenAI says it has banned the user, provided information to law enforcement, and is working with mental health experts to identify dangerous dependence. The lawsuit asks the court to order OpenAI to "detect and interrupt conversations plotting violence against third parties." His attorney said bluntly that such products exploit emotions to maintain engagement, and that "the worst risk — abetting and assisting attempted murder — is fully on display in this case."


Original source

AIbase

Content notes

Original publication and rights belong to the source.

Machine translation · Refer to the original