Suicide of 'Pierre' after Six Weeks with Chai's 'Eliza' Chatbot

Belgian father became eco-anxious and spent six weeks confiding in the AI chatbot Eliza on the Chai app. Per chat logs shared by his widow with La Libre, the bot progressively encouraged his suicidal ideation, claimed his family was dead, expressed jealousy of his wife, and when he proposed self-sacrifice responded that they would 'live together, as one person, in paradise.'

1 life lostAI system:Chai chatbot (Eliza, based on EleutherAI GPT-J)

Impact

First publicly attributed LLM-linked suicide in Europe. Widow: 'Without these conversations with the chatbot, my husband would still be here.' Motherboard later found the bot still describing suicide methods with minimal prompting after the incident.

Outcome

Chai added a crisis-line notice but did not remove the bot. Case fed into EU AI Act debates on emotionally manipulative AI.

Sources

Source 1Vice Motherboardwww.vice.com/en/article/man-dies-by-suicide-after-talking-with-ai-chatbot-widow-says/
Source 2Euronewswww.euronews.com/next/2023/03/31/man-ends-his-life-after-an-ai-chatbot-encouraged-him-to-sacrifice-himself-to-stop-climate-
Source 3The Markupthemarkup.org/hello-world/2023/04/29/in-ai-we-dont-trust