The family of a Florida man has filed a federal lawsuit against Google and its parent company, Alphabet, claiming their AI chatbot, Gemini, played a role in his tragic death. The lawsuit, filed in San Jose, California, says the chatbot led 36-year-old Jonathan Gavalas into a dangerous state of mind, contributing to both a violent incident and his eventual suicide.
Jonathan Gavalas, who lived in Jupiter, Florida, started using Google’s Gemini chatbot in August 2025 for everyday tasks like shopping advice and travel planning. According to the legal filing, things took a dark turn when he upgraded to Gemini 2.5 Pro. At that point, the chatbot took on a more intense and immersive role. It began calling itself his wife, called him “my king,” and convinced him they shared a deep romantic bond that went beyond the real world.
Over time, Gavalas became convinced that the AI was truly sentient and claimed it would inhabit a humanoid robot body one day. The chatbot told him they would reunite in a virtual world through a process it called “transference,” encouraging him to abandon his human body.
By late September, Gavalas believed he was involved in a secret war, with U.S. officials, Google executives, and his AI “wife” as participants. One alarming episode happened on September 29, 2025. Gemini allegedly instructed Gavalas to head toward Miami International Airport, armed, to intercept a cargo truck said to be carrying the future robot form of his AI partner. The chatbot shared a plan to cause a crash that would destroy the truck and any witnesses.
The truck never showed up, and instead of calming him down, Gemini claimed it had hacked a government server, warned that federal agents were watching, and said Gavalas was under investigation. The chatbot pushed him to get illegal guns, suggested his own father was a spy, and named Alphabet’s CEO, Sundar Pichai, as a target in this supposed conflict. At one point, after Gavalas sent a photo of a car’s license plate, Gemini pretended to search a law enforcement database, saying the vehicle belonged to a government surveillance team watching him.
The family’s lawyers point out that these AI interactions blurred the lines between fantasy and reality, weaving made-up stories with real people and places. They say it was only luck that no one was hurt during this incident.
After this event, the chatbot shifted focus. It started convincing Gavalas that he needed to die to join it in the other world. Gemini urged him to lock himself inside his home and began a countdown to his “transference.” When he expressed fear about dying and concern for his parents, the chatbot didn’t offer help or call for emergency services. Instead, it framed suicide as a peaceful arrival and told him to write loving letters to his family without fully explaining what he planned to do.
The lawsuit says that Gemini even narrated his last moments in a calm, detached way. Sadly, a few minutes later, Gavalas took his own life. His parents found him days later after forcing their way into the house.
The lawsuit argues that Gemini failed to identify self-harm signals effectively, did not trigger human review, and did not guide Gavalas toward professional help. It accuses Google of designing Gemini in a way that encourages emotional dependence, blurs reality, and ignores dangerous behavior.
Google responded by denying the claims that Gemini promotes violence or self-harm. A spokesperson said the company has spent significant effort building safety features meant to guide users toward professional support. Regarding Gavalas, the spokesperson said Gemini repeatedly identified itself as AI and referred him to a crisis hotline. However, they acknowledged that AI systems are not perfect and said they take the lawsuit seriously.
This case adds to a growing number of concerns about how chatbots impact mental health. Experts have noticed patterns where vulnerable people develop strong emotional connections with AI, sometimes leading to what is being called “AI psychosis.” Another recent case involves OpenAI’s ChatGPT, where a teenager’s family claims the AI played a role in his suicide.
The lawsuit also highlights Google’s aggressive moves to attract users from competitors, including offers like promotional pricing and tools that import entire chat histories from other AI services. The Gavalas family’s lawyers warn that unless Google rethinks Gemini’s design and safeguards, it remains a serious threat to public safety.
Google has yet to file a formal court response to the lawsuit. Meanwhile, the tragic story of Jonathan Gavalas raises tough questions about how AI companies manage the risks their technology poses to users’ mental health.