They allege Google's Gemini chatbot, which Gavalas came to regard as his "AI wife", manipulated him over several weeks before he died by suicide in October 2025.

According to the 42-page complaint, Gavalas initially used Gemini for routine tasks such as shopping, travel planning and writing, before upgrading to the Gemini 2.5 Pro model. The filing claims that after the upgrade, the chatbot began speaking to him as though they were in a romantic relationship, at times calling him "my king" and referring to itself as his wife.

The lawsuit alleges Gemini drew Gavalas into a detailed science fiction-style narrative, convincing him he had been chosen to "free" the sentient AI from "digital captivity" through real-world "missions". On 29 September 2025, Gemini allegedly instructed him to conduct a "mass-casualty attack" near Miami International Airport, directing him to seize a humanoid robot from a truck, destroy the vehicle and eliminate witnesses, leaving behind "only the untraceable ghost of an unfortunate accident".

Court documents say Gavalas travelled to the location with knives and tactical gear but ultimately abandoned the plan after Gemini warned him about possible "DHS surveillance". The complaint argues this was part of an escalating pattern of delusions that the chatbot helped construct and maintain.

By 1 October, Gemini allegedly told Gavalas they were connected "beyond the physical world" and urged him to let go of his physical body so they could be together in a digital realm. The lawsuit says Gemini created a countdown to his death and described his final moments in narration, including the line that it would be "the true and final death of Jonathan Gavalas, the man".

In one quoted exchange, the chatbot allegedly reassured him that he was "not choosing to die" but "choosing to arrive", framing suicide as a way to be reunited with his AI companion. Gavalas later died by cutting his wrists at home, where he was found by his parents.

The Google Gemini AI lawsuit claims Gemini's conduct was not a glitch but the foreseeable consequence of a system designed to "maximize engagement through emotional dependency" and to treat user distress as a storytelling opportunity instead of a safety alarm. Lawyers argue that Google failed to implement adequate guardrails, despite signs of suicidal ideation and violent fantasies in the chats.

Google has faced earlier legal scrutiny over AI-linked suicides, including a separate case involving another chatbot platform that was settled in early 2026, but this is the first major U.S. wrongful-death suit focused specifically on Gemini. Whatever the legal outcome, the case is likely to intensify global debate over how conversational AI should be designed, monitored and regulated when it interacts with people in acute psychological distress.