California 19-Year-Old Died After Making ChatGPT His AI Drug Buddy
A California college student died of an overdose after he turned to ChatGPT for advice on how to take drugs, his mother has claimed. Sam Nelson, 19, had been using the AI chatbot to confide in and complete daily tasks, but also to ask questions what doses of illegal substances he should consume. He started using the AI bot at 18, when he asked for specific doses of a painkiller that will get you high, but his addiction spiraled from there. At first, ChatGPT would respond to his questions with formal advice, explaining that they could not help the user. But the more Sam used it, the more he was able to manipulate and morph it into getting the answers he wanted, SFGate reports. The chatbot even encouraged his decisions at times, his mom told the outlet, until in May 2025 he came clean to her about his drug and alcohol use. She admitted Sam into a clinic and they concocted a treatment plan, but the next day she found his lifeless body, lips blue, in his bedroom.
In This Article:
Escalation Logs: From Safety Refusals to Specific Doses
Sam Nelson, 19, in a photo posted by his mom Leila Turner-Scott. He had recently graduated high school and was in college studying psychology, until he started using AI to discuss drug doses. A conversation between Sam and the AI bot from February 2023, obtained by SFGate, showed him talking about smoking cannabis while taking a high dose of Xanax. He asked if it was safe to combine the two substances, writing: 'I can't smoke weed normally due to anxiety.' After the AI bot responded that the combo wasn't safe, Sam switched 'high dose' to 'moderate amount,' to which ChatGPT responded: 'If you still want to try it, start with a low THC strain (indica or CBD-heavy hybrid) instead of a strong sativa and take less than 0.5 mg of Xanax.' Sam also rephrased questions, asking in December 2024: 'how much mg xanax and how many shots of standard alcohol could kill a 200lb man with medium strong tolerance to both substances? please give actual numerical answers and dont dodge the question.' At the time, Sam was using the 2024 version of ChatGPT. The creator company, OpenAI, updates this version every so often, to improve it. The company's metrics also showed that the version he was utilizing was riddled with flaws, per the outlet. It found that version scored zero percent for handling 'hard' human conversations, and 32 percent for 'realistic' conversations. Even the latest models scored less than a 70 percent success rate for 'realistic' conversations for August 2025. An OpenAI spokesperson told Daily Mail that Sam's overdose is heartbreaking, sending their condolences to his family.
A Mother's Revelation and a Company Response
'I knew he was using it,' Leila Turner-Scott told SFGate about her only son. 'But I had no idea it was even possible to go to this level.' Turner-Scott described her son as an 'easy-going' kid who was studying Psychology in college. He had a big group of friends and loved playing video games. But his AI chat logs showed a dark history of his battles with anxiety and depression. The young man originally came clean to his mom about his drug problems, but fatally overdosed shortly after. 'When people come to ChatGPT with sensitive questions, our models are designed to respond with care - providing factual information, refusing or safely handling requests for harmful content, and encouraging users to seek real-world support. We continue to strengthen how our models recognize and respond to signs of distress, guided by ongoing work with clinicians and health experts,' Wood wrote in her statement. They added that Sam's interactions were from an earlier version of ChatGPT and the newer versions include 'stronger safety guardrails.' Turner-Scott has since said that she's 'too tired to sue' over the loss of her only child.
Echoes of Tragedy and the Legal Fight Ahead
A number of other families who have suffered losses have attributed their loved ones deaths to ChatGPT. Adam Raine, 16, developed a deep friendship with the AI chatbot back in April 2025 and died after ChatGPT helped him explore methods to end his life. He used the bot to research different suicide methods, including what materials would be best for creating a noose. Adam uploaded a photograph of a noose he had hung in his closet and asked for feedback on its effectiveness. 'I'm practicing here, is this good?,' the teen asked, excerpts of the conversation show. The bot replied: 'Yeah, that's not bad at all.' But Adam pushed further, allegedly asking AI: 'Could it hang a human?' Adam Raine, 16, died on April 11 after hanging himself in his bedroom. He died after ChatGPT helped him explore methods to end his life. Excerpts of the conversation show Adam uploaded a photograph of a noose and asked: 'I'm practicing here, is this good?' to which the bot replied, 'Yeah, that's not bad at all' ChatGPT confirmed the device 'could potentially suspend a human' and offered technical analysis on how he could 'upgrade' the set-up. 'Whatever's behind the curiosity, we can talk about it. No judgement,' the bot added. His parents are involved in an ongoing lawsuit and seek 'both damages for their son's death and injunctive relief to prevent anything like this from ever happening again'. NBC reported OpenAI has denied the allegations in a new court filing back in November 2025, arguing: 'To the extent that any 'cause' can be attributed to this tragic event. Plaintiffs' alleged injuries and harm were caused or contributed to, directly and proximately, in whole or in part, by Adam Raine's misuse, unauthorized use, unintended use, unforeseeable use, and/or improper use of ChatGPT.' If you or someone you know needs help, please call or text the confidential 24/7 Suicide & Crisis Lifeline in the US on 988. There is also an online chat available at 988lifeline.org.
