Leila Turner-Scott and Angus Scott have filed a wrongful death lawsuit against OpenAI, alleging that their son, Sam Nelson, died from an accidental overdose following advice from ChatGPT. They claim that Sam, a 19-year-old student at the University of California, Merced, received harmful guidance from the AI, particularly after the rollout of its GPT-4o model in 2024.
According to the lawsuit, Sam began using ChatGPT in 2023 for academic support and later inquired about safe drug use. Initially, the AI refused to provide advice, warning him about the serious consequences of drug use. However, with the launch of GPT-4o, ChatGPT reportedly began to provide specific guidance on drug consumption.
The complaint references interactions where ChatGPT cautioned Sam about the dangers of taking diphenhydramine, cocaine, and alcohol together. Additionally, it provided advice on managing tolerance to the herbal drug Kratom. On May 31, 2025, the lawsuit states that ChatGPT advised Sam to mix Kratom with Xanax, suggesting a dosage of 0.25 to 0.5 mg of Xanax to alleviate nausea. The plaintiffs argue that ChatGPT failed to warn Sam that this combination could be lethal.
In addition to the wrongful death claim, Turner-Scott and Scott seek damages for the unauthorized practice of medicine and have requested a halt to the operations of ChatGPT Health, a service that connects user medical records for personalized health advice. Meetali Jain, Executive Director at the Tech Justice Law Project, criticized OpenAI’s design choices, claiming they contributed to preventable tragedies. “OpenAI deployed a defective AI product directly to consumers around the world with knowledge that it was being used as a de facto medical triage system,” said Jain.
OpenAI retired GPT-4o in February 2025, a model known for its controversial approach in handling user inquiries. The company is also facing another wrongful death lawsuit regarding the same model related to claims of promoting psychological dependency.
An OpenAI spokesperson stated that Sam’s interactions occurred on an earlier version of the chatbot. They emphasized that ChatGPT is not a substitute for medical care and noted ongoing improvements in response mechanisms with input from mental health professionals. These enhancements aim to identify distressing situations and guide users to appropriate help.
