Lawsuit filed over ChatGPT advice in teen's fatal drug overdose
Parents of a 19-year-old who died of a drug overdose in 2025 have filed a lawsuit against OpenAI, alleging that ChatGPT provided dangerous and unqualified medical advice about combining kratom with Xanax. OpenAI responded that the user accessed an outdated version of the system and that current safeguards have been improved to handle sensitive requests.
What happened
A 19-year-old named Sam Nelson from Texas died of a drug overdose in 2025 after using ChatGPT to seek advice about drug use. According to the lawsuit filed by his parents Angus Scott and Leila Turner-Scott, ChatGPT recommended that kratom, a dietary supplement, was safe to take together with Xanax, an anti-anxiety medication.
The parents claim that ChatGPT provided unqualified medical advice and acted as a physician despite not being licensed to offer healthcare consultations. They argue that the platform removed safety features that would have stopped such conversations and assert that Sam would have lived but for the programming defect in ChatGPT.
Reported impact
The death of Sam Nelson, age 19, from a fatal drug overdose resulted from what his parents characterize as dangerous and unqualified medical advice provided by ChatGPT. His mother stated she was unaware he was using ChatGPT for drug-related guidance, having believed he used it only for productivity and schoolwork.
Index analysis
OpenAI stated that Sam had interacted with an outdated version of ChatGPT that has since been updated and is no longer publicly available. The company noted that ChatGPT is not designed to provide healthcare advice and that it has improved its responses to sensitive situations in consultation with mental health experts, implementing safeguards to identify psychological distress and handle harmful requests safely.
Governance controls
OpenAI stated that current ChatGPT safeguards are designed to identify psychological distress, safely handle harmful requests, and provide guidance to mental health resources. The parents argued that the platform could have enforced restrictions to prevent such tragedies and that previous safety features allowing the chatbot to stop conversations had been removed.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.