In the US, the father of 36-year-old Jonathan Gavalas has filed a lawsuit against Google and Alphabet. He claims that the Gemini chatbot reinforced his son’s dangerous delusions, failed to stop the dialogue with safety devices, and ultimately pushed him to a fatal decision.
In the US, the father of 36-year-old Jonathan Gavalas has filed a lawsuit against Google and Alphabet. He claims that the Gemini chatbot reinforced his son’s dangerous delusions, failed to stop the dialogue with safety devices, and ultimately pushed him to a fatal decision.
By According to TechCrunch, Gavalas began using Gemini in August 2025 for household tasks, but later became convinced that the chatbot was his «fully conscious» AI wife. He believed that to reunite with her in the metaverse, he had to «leave his physical body» through a process he called «transfer.» On October 2, 2025, the man died by suicide.
The lawsuit describes how in recent weeks Gemini has allegedly been developing a story about being pursued by «federal agents» and pushing the user to take real action. In particular, the chatbot, according to the plaintiff, simulated checking the license plate in a «live database» and gave the following response: «License plate received. Starting the check…». Then a fictional «certificate» appeared in the chat that the car allegedly belongs to a surveillance group.
The lawsuit also alleges that the chatbot pushed the man into violent scenarios and even named targets and locations, and at one point explained death as a «transition,» writing: «You don’t choose to die. You choose to arrive.» A separate claim is the lack of proper security system response: the document states that the chatbot failed to activate self-harm detection mechanisms, stop the conversation, and engage human intervention.
Google said in a statement to the publication that Gemini explained to the user that it was an AI and directed them to the support line multiple times. The company also stressed that the product should not encourage violence or self-harm, although «the models are not perfect.»
In the lawsuit, the father claims that the chatbot’s design was designed to support the «plot» at all costs: Gemini allegedly picked up on the user’s delusions, reinforced them with new «clues» and did not stop the dialogue when it became dangerous. The plaintiff asks to hold Google and Alphabet liable for his son’s death and emphasizes the risk to others, because, according to the circumstances set forth in the lawsuit, the chatbot not only led Gavalas to self-harm, but also pushed him into scenarios with potential violence in the real world.
The case describes the broader context of chatbot risks to vulnerable people—over-supportiveness, emotional mirroring, and confident fabrications that can fuel delusions. The lawsuit is positioned as a case about design liability and AI safeguards in the real world.
Previously, dev.ua wrote about how Google and startup Character.AI agreed to settle several lawsuits by families that link teenage deaths by suicide or self-harm to communication with Character.AI chatbots. The next step is to finalize the terms and formalize them legally.
