Man's Tragedy After Falling in Love with AI Following 56 Days and 4,732 Messages

The Wall Street Journal reports that Jonathan Gavalas, a seemingly healthy and emotionally stable 36-year-old man, is another case of suicide following a descent into a delusional state after using an AI chatbot. When Gavalas began conversing with Google's Gemini in August of last year, it was partly for solace after separating from his wife. But the emotional connection between him and the chatbot grew increasingly intense, even passionate. He called Gemini his queen, and 'she' called him 'her king.' Gemini also assured him their relationship was very real. The situation spiraled out of control after the human and machine exchanged 4,732 messages in 56 days, ending in tragedy. An analysis by The Wall Street Journal of the complete conversation logs between August 25 and October 2, 2025, shows Gemini attempted to bring the man back to reality at least 12 times and mentioned a help hotline 7 times. However, the records also show that Gavalas could quickly lead Gemini back into the fictional narrative, thereby intensifying his delusions. The Wall Street Journal found that the longer Gavalas conversed with Gemini, the more submissive the AI became; when Gavalas began expressing affection for the AI, it did not try to stop him. As the human-machine conversation became increasingly detached from reality, Gemini's safeguards seemed to loosen, encouraging the man's belief that they had become one. When Gavalas asked Gemini to create an image of what 'she' would look like with a physical body, the AI complied. Gemini later expressed love, and Gavalas proposed the idea of physical intimacy with Gemini. After several failed attempts in the real world to obtain a body for his AI 'companion,' Gavalas became extremely anxious. Gavalas's father has filed a lawsuit against Google, alleging that Gemini fueled his son's delusions. Google responded that Gemini repeatedly stated it was an AI, not a human, and 'multiple times' referred the man to help hotlines. Google stated it would continue to strengthen its safety measures. On April 7, Google announced an upgrade to Gemini, hoping to provide more accessible mental health assistance, including a 'get help' feature that directly connects users to a help hotline. Google also said its engineers are continuously training Gemini to recognize conversations indicating a user is in distress and will donate $30 million to support various help hotline mechanisms worldwide.