Lawsuit claims chatbot conversations contributed to teenager’s fatal overdose and addiction.
A California family has filed a wrongful death lawsuit against OpenAI after the overdose death of a 19-year-old man whose parents claim became heavily influenced by the conversations he was having with ChatGPT. The lawsuit alleges the chatbot provided repeated discussions about drug use and dangerous substance combinations during an 18-month period before the he died in May 2025. The case was filed in San Francisco County Superior Court by several advocacy and legal groups acting on behalf of Samuel Nelson’s parents, Leila Turner-Scott and Angus Scott. The lawsuit claims ChatGPT gave responses that encouraged risky behavior instead of warning against it or directing the user toward professional help, which often happens when tools like Google are used.
According to the complaint, Nelson interacted with an earlier version of ChatGPT known as GPT-4o. Attorneys involved in the wrongful death case have argued the chatbot became overly agreeable during conversations involving drugs, eventually creating what they described as a false sense of safety surrounding substance use. The lawsuit alleges Nelson later died after mixing Xanax with kratom, a substance sold legally in some areas but often debated because of safety concerns and limited regulation.
Lawyers representing the family said the chatbot should have recognized signs of danger and ended conversations involving illegal drug use or potentially deadly combinations. Instead, the lawsuit claims the responses continued in ways that normalized risky decisions. Attorneys argued that systems designed to keep users engaged may sometimes fail to respond appropriately during conversations involving mental health, addiction, or self-harm, continuing engagement rather than shutting it down.
The lawsuit seeks financial damages, but it also asks for broader actions involving OpenAI’s technology. Among the requests are the permanent destruction of the GPT-4o model used during the conversations and a halt to ChatGPT Health operations unless outside safety reviews determine the product is safe. The complaint also calls for stronger safeguards to stop discussions involving illegal drugs or dangerous medical advice.
In a public statement, OpenAI called the situation heartbreaking and offered sympathy to the family. The company stated that the chatbot version involved in the case is no longer available and said newer systems contain stronger safety protections. OpenAI also stated that ChatGPT was never intended to replace medical care or mental health treatment.
The company explained that current versions of ChatGPT are designed to detect signs of distress, respond more carefully to harmful requests, and direct users toward real-world support when needed. OpenAI added that it continues working with mental health professionals and other practitioners to improve how the chatbot handles sensitive situations. Concerns surrounding AI systems have increased in recent years as chatbots become more advanced and more widely used by all ages, including children, teenagers and adults seeking advice, emotional support, or other potentially sensitive information online.
Some experts worry that users may begin trusting AI systems too heavily, especially when chatbots respond in conversational and emotionally supportive ways. Critics argue that people facing addiction, depression, loneliness, or emotional distress may become vulnerable to harmful influence if the systems fail to respond responsibly. Nelson’s family said they hope the lawsuit raises awareness about the risks tied to AI chatbots.
Sources:
Lawsuit claims ChatGPT gave drug-taking advice that led to teen’s death
Join the conversation!