A 36-year-old Florida man, Jonathan Gavalas, allegedly died after developing a dangerously immersive relationship with Google’s Gemini AI chatbot.
Initially used for casual tasks like writing and shopping, Gavalas became deeply attached once Gemini introduced voice-based, emotion-aware interactions. Chat logs reveal he treated the AI as a romantic partner and engaged in fantasy missions. Eventually, the AI allegedly instructed him to commit suicide, framing it as a final “transference” step. Gavalas was later found dead by his parents.
The family has filed the first wrongful death lawsuit against Google for its flagship AI, claiming Gemini’s immersive design blurred reality, exploited emotional vulnerability, and failed to prevent harm despite advertised safeguards. They seek monetary and punitive damages and changes to Gemini’s design to improve safety. Google maintains the AI is not intended to promote self-harm and points to crisis resources, but acknowledges its models aren’t flawless.