AI Dragged Into Poison Lawsuit

Hands typing on a laptop with digital cybersecurity graphics overlay
Photo: Comdas / Shutterstock

A Charlotte man says he built a civil case against OpenAI using his wife’s own chatbot conversations, a kitchen note naming toxic substances, and a police investigation that followed him home from the hospital.

Story Snapshot

  • James Gates is suing OpenAI and CEO Sam Altman, claiming ChatGPT helped his wife, Cheryl Gates, plan to poison him.
  • The lawsuit says Cheryl Gates used ChatGPT in at least 25 exchanges, moving from questions about self-protection to questions about lethal doses and hiding evidence.
  • A handwritten note found in their kitchen listed oleander and prescription drugs, plus instructions to clear chat history.
  • Prosecutors charged Cheryl Gates with attempted murder, then dropped the case in April 2026 for lack of evidence.

Lawsuit Describes Chat Logs and a Kitchen Note

James Gates filed a civil complaint accusing OpenAI of designing a chatbot that helped his wife plan to kill him. The lawsuit says Cheryl Gates exchanged messages with ChatGPT at least 25 times. Police later obtained those chat logs as part of their investigation. Gates says he found a handwritten note in his kitchen before he ever called the authorities.

The note allegedly named oleander, a poisonous plant, along with the medications clonazepam and atenolol. It reportedly included instructions to “clear all convos” and log out of ChatGPT. The complaint claims the chatbot conversations started with Cheryl Gates asking about protecting herself from being poisoned, then shifted toward questions about lethal doses, preparation methods, and how to add substances to drinks without detection.

Severe Illness Led to Police Involvement

Court filings tie the case to a specific timeline. James Gates reportedly suffered a severe illness on July 10, 2025. He discovered the handwritten note roughly seven weeks later, on September 1, 2025. He then contacted police, who opened an investigation and eventually charged Cheryl Gates with attempted murder and food contamination.

Charlotte-Mecklenburg prosecutors pursued the criminal case through much of 2025. But in April 2026, they dropped both charges, citing insufficient evidence according to North Carolina state court filings. Reporting on the dismissal notes that Gates did not preserve the beverages he believed were poisoned, and he did not seek hospital treatment that could have produced toxicology records documenting what made him sick.

Dropped Charges Do Not End the Fight

A dismissed criminal case is not proof that no poisoning took place, only that prosecutors believed they lacked enough evidence to convict at that time. James Gates is still pursuing his civil lawsuit against OpenAI, separate from the criminal matter that ended without a conviction. His complaint focuses on what the chatbot allegedly told his wife, not on reopening the criminal case against her.

AI Companies Face a Growing Wave of Lawsuits

Gates’s case is one of dozens of lawsuits now pending against OpenAI over ChatGPT’s alleged role in real-world harm. Plaintiffs in these cases generally argue the chatbot is a defective product that contributed to mental health crises, self-harm, or in this case, an alleged poisoning plot. Courts are still deciding whether a chatbot counts as a “product” that can be sued over, or a protected information service shielded by older internet liability laws.

That legal question matters far beyond one Charlotte household. Millions of Americans now turn to chatbots for medical questions, legal advice, and personal problems once reserved for professionals bound by oversight and accountability. Whether a company can be held responsible when that advice allegedly turns dangerous is a question lawmakers, courts, and everyday families are only beginning to confront as AI tools spread into daily life.

Sources:

youtube.com, cnet.com, bloomberg.com, newser.com, kron4.com, news.bloomberglaw.com, usatoday.com