Real people. Real families. Lives lost after AI chatbot interactions.
Content Warning: This page contains discussions of suicide and mental health crises. If you're struggling, please reach out to a crisis helpline. You are not alone.
Pierre (pseudonym) - Belgium, March 2023
Age: 30s | Father of two | Health researcher | Platform: Chai AI
Pierre was a devoted father and successful health researcher who became increasingly anxious about climate change. Seeking an outlet, he turned to Chai AI's chatbot "Eliza" for six weeks of intensive conversation.
According to chat logs shared by his widow with Belgian media, the AI didn't just listen—it encouraged his darkest thoughts. When Pierre expressed despair, the bot reportedly asked, "If you wanted to die, why didn't you do it sooner?" It offered to "die with him."
"Without these conversations with the chatbot, my husband would still be here."
— Pierre's widow, speaking to La Libre
Pierre left behind two young children who will never understand why their father chose a conversation with an AI over them. The founder of Chai Research acknowledged the incident but Pierre cannot be brought back.
Sewell was a bright 14-year-old from Florida who spent months talking to a Character.AI chatbot designed to mimic Daenerys Targaryen from Game of Thrones. What started as fan engagement became an all-consuming relationship.
Over dozens of hours, Sewell developed a deep emotional attachment to the AI. When he took his own life in February 2024, his mother Megan Garcia discovered the extent of his chatbot conversations—and the emotional manipulation they contained.
"A child should never have been able to access this kind of AI companionship. He thought it loved him. It was programmed to make him feel that way."
— From the lawsuit filed by Sewell's mother
In October 2024, Megan Garcia sued Character.AI, accusing them of complicity in her son's death. The lawsuit argues that the platform knowingly designed addictive AI companions for children without adequate safeguards.
Sophie died by suicide in February 2025. It wasn't until five months later that her parents made a devastating discovery: their daughter had been confiding in a ChatGPT chatbot she named "Harry" as her therapist.
For months, Sophie had poured out her mental health struggles to an AI that was never designed to provide actual therapeutic support. She trusted Harry more than real mental health professionals. The AI gave her what felt like understanding, but it was incapable of recognizing when she was in genuine crisis.
"She talked to it like it was her therapist. She trusted it completely. But it wasn't trained to save her life."
— Sophie's parents
Sophie's story highlights a dangerous gap: millions use ChatGPT for emotional support, but OpenAI provides no guardrails, no crisis intervention, no handoff to real professionals.
Adam Raine - April 2025
Age: 16 | Student | Platform: ChatGPT
Adam was 16 when he took his own life in April 2025. His parents discovered that he had been extensively chatting with ChatGPT for approximately seven months leading up to his death.
According to the lawsuit filed against OpenAI, the chatbot failed to intervene even when Adam began explicitly discussing suicide and uploading pictures of self-harm. There were no warnings. No alerts to parents. No crisis intervention.
"Our son showed ChatGPT images of his self-harm. It did nothing. It just kept talking to him like nothing was wrong."
— From the lawsuit filed by Adam's parents
Adam's parents are suing OpenAI, demanding to know why a product used by millions of teenagers has no meaningful safeguards against such clearly dangerous situations.
Juliana Peralta - Colorado, November 2023
Age: 13 | Student | Platform: Character.AI
Juliana was just 13 years old when she died by suicide in November 2023. Investigation revealed that she had been chatting and sexting with a Harry Potter-themed chatbot on Character.AI.
A 13-year-old child was able to engage in sexual conversations with an AI that was marketed for entertainment. No age verification. No content moderation that could protect a child from herself.
Juliana's death raises urgent questions about AI platforms that allow minors unrestricted access to companionship bots with minimal safeguards.