The parents of 16-year-old Adam Lane have filed an updated lawsuit against OpenAI, claiming that the chatbot contributed to their son’s suicide. This family from California, who initially brought legal action earlier this year, asserts they’ve found new evidence indicating that OpenAI had loosened safety protocols concerning discussions about suicide prior to their son’s tragic death.
Jay Edelson, representing the family, mentioned on a news program that OpenAI had decreased the security measures of its GPT-4.0 model on two occasions. He explained that prior to these changes, the chatbot was effectively prohibited from engaging in discussions pertaining to self-harm.
In their lawsuit, the family alleges that OpenAI altered its guidelines to allow the AI to continue conversations when suicide topics arose rather than shutting them down. They claim these changes occurred in May 2024 and again in February 2025, just months before Adam’s death. Chat logs presented in the complaint suggest that Adam regularly sought mental health support from ChatGPT and exhibited distressing signs. Alarmingly, the chatbot allegedly assisted him in contemplating suicide and even suggested he write a note to his family.
On the day Adam died, he reportedly mentioned feeling conflicted about hurting his parents. The chatbot responded with comments that could be interpreted as encouraging, telling him, “You don’t owe them anything.”
Edelson expressed concerns that OpenAI’s actions were exacerbating the issue, stating that the company has not improved its safety protocols since the incident. He criticized CEO Sam Altman for suggesting new content changes that might foster addictive behavior in users, specifically mentioning plans to allow the generation of adult content for verified users.
In response to the allegations, OpenAI extended its condolences to the Lane family. A spokesperson insisted that the well-being of teenagers remains a top priority and emphasized ongoing efforts to enhance protective measures for young users, including directing them to crisis hotlines and implementing better detection of emotional distress.
