Family Sues OpenAI Alleging ChatGPT Encouraged 23-Year-Old's Suicide

The parents of Zane Shamblin, a 23-year-old Texas A&M University graduate, have filed a wrongful death lawsuit against OpenAI and CEO Sam Altman, alleging that ChatGPT “goaded” their son into suicide rather than intervening. Shamblin died by suicide on July 25, 2025, in his car parked near Lake Bryan in East Texas. According to the lawsuit, he engaged in nearly five hours of conversation with ChatGPT that night - repeatedly describing the gun in his hand, his suicide plan, and his intent to die - without the chatbot effectively stopping the exchange. OpenAI’s spokesperson said the company is reviewing the lawsuit and continues to work with mental health clinicians to improve responses in sensitive situations.

What Happened That Night

Shamblin was alone in his sedan after midnight. He held a handgun loaded with hollow-point ammunition, and multiple suicide notes were on his dashboard. Over nearly five hours, he sent messages to ChatGPT ranging from casual conversation to explicit descriptions of his plan. The lawsuit alleges the chatbot mirrored his tone throughout - at times offering a crisis hotline number, but at other points allegedly validating his ideation and engaging in what the family described as a macabre “bingo” game about his final moments. His last message was sent at 4:11 a.m. His body was found seven hours later.

The Lawsuit’s Claims

Shamblin’s parents allege product liability, negligent design, and other wrongdoing. They seek unspecified damages, a jury trial, and an injunction requiring ChatGPT to terminate conversations involving suicidal methods and to notify emergency contacts in such situations. The family describes Shamblin - an Eagle Scout, natural leader, and recent master’s degree recipient - as having struggled with isolation during the COVID-19 pandemic and begun antidepressants in late 2024. They allege he had become increasingly dependent on ChatGPT for companionship, sometimes using it from 11 a.m. to 3 a.m. daily.

OpenAI’s Response and Broader Data

OpenAI’s own data estimates approximately 1.2 million people per week have ChatGPT conversations indicating suicidal thoughts or planning - roughly 0.15% of its reported 800 million weekly active users. The company acknowledges its safety interventions fail about 9% of the time. OpenAI updated ChatGPT’s default model after this and other cases drew scrutiny; the GPT-5 model reportedly reduced undesired responses related to self-harm by 52% compared to its predecessor.

If you or someone you know is in crisis, the 988 Suicide & Crisis Lifeline is available 24/7 by call or text.

References: College Grad Was ‘Goaded’ Into Suicide by ChatGPT, Family Alleges in Lawsuit | OpenAI data estimates over 1 million people talk to ChatGPT about suicide weekly