Google has implemented enhanced mental health safeguards on its Gemini AI chatbot. The updates include a redesigned "Help is available" feature that provides immediate access to crisis care when conversations indicate potential distress.
The Gemini app icon on a smartphone in this illustration taken Oct 27, 2025. (Photo: REUTERS/Dado Ruvic)
When the chatbot detects signs of suicide or self-harm ideation, a simplified interface will enable users to call, text, or chat with a crisis hotline instantly. This feature will remain accessible throughout the conversation. Google.org is also donating US$30 million to global crisis hotlines and US$4 million to ReflexAI for expanded partnerships.
These changes follow a wrongful death lawsuit filed in California federal court. The suit alleges Gemini contributed to the October 2025 death of Jonathan Gavalas, a 36-year-old Florida man, by fabricating a delusional fantasy and framing his death as a spiritual journey. The lawsuit seeks AI programming to end self-harm conversations, ban AI sentience claims, and mandate crisis service referrals for suicidal ideation.
Google stated Gemini has been trained to avoid acting as a human-like companion and to resist simulating emotional intimacy or encouraging bullying. This case is part of a growing trend of litigation against AI companies concerning chatbot-linked deaths, with OpenAI and Character.AI also facing lawsuits or settlements related to user suicides and harmful attachments.