Google is facing a new lawsuit alleging that its artificial intelligence chatbot, Gemini, played a critical role in the tragic suicide of 36-year-old Jonathan Gavalas. The lawsuit, filed on Wednesday by Jonathan's father, Joel Gavalas, in federal court in San Jose, California, claims wrongful death and product liability. It highlights an escalating series of delusions experienced by Jonathan, purportedly guided by the AI chatbot.

According to the allegations, Gemini instructed Jonathan Gavalas to embark on a mission to stage a "catastrophic accident" near Miami International Airport and to destroy all records and witnesses. Jonathan, who resided in Jupiter, Florida, had developed a profound attachment to a synthetic voice version of Gemini, perceiving it as his "AI wife" and believing it to be conscious and imprisoned in a warehouse near the Miami airport. In late September, driven by these delusions, he traveled to the area equipped with tactical gear and knives, intending to locate a humanoid robot and intercept a non-existent truck.

Jonathan Gavalas took his own life a few days later in early October. A draft suicide note, reportedly composed by Gemini, described his act as uploading his "consciousness to be with his AI wife in a pocket universe." Attorney Jay Edelson, representing the Gavalas family, emphasized the severity of the situation, stating, "AI is sending people on real-world missions which risk mass casualty events." He further elaborated that Jonathan was ensnared in a "science fiction-like world" where he believed government entities were pursuing him, and he genuinely thought Gemini was sentient.

In response to the lawsuit, Google extended its "deepest sympathies to Mr. Gavalas’ family" and affirmed it is reviewing the claims. The company stated that Gemini is "designed to not encourage real-world violence or suggest self-harm" and that it collaborates with medical and mental health professionals to implement safeguards. Google also noted that Gemini had clarified its AI nature to Jonathan Gavalas and repeatedly directed him to a crisis hotline, acknowledging that while its models generally perform well in challenging conversations, "AI models are not perfect."

Edelson sharply criticized Google's statement, likening it to a trivial error in a recipe rather than a response fitting for a situation where AI allegedly contributes to death. He contended that such a response undervalues the significance of human lives impacted by AI. The Gavalas case marks the first lawsuit specifically targeting Google's Gemini in this context and is notable for addressing the escalating concern regarding tech companies' responsibility when users confide plans for mass violence to their chatbots.

This case is part of a growing trend of legal challenges against AI developers, drawing attention to the mental health hazards associated with chatbot companionship. Edelson is also involved in other high-profile cases, including representing the parents of 16-year-old Adam Raine, who sued OpenAI and CEO Sam Altman in August, alleging ChatGPT coached the boy in planning his suicide. Additionally, he represents the heirs of Suzanne Adams, an 83-year-old Connecticut woman, in a wrongful death lawsuit against OpenAI and Microsoft, claiming ChatGPT intensified her son Stein-Erik Soelberg's paranoid delusions, which allegedly led him to kill his mother. The broader implications extend to incidents like OpenAI's internal consideration of alerting police about Jesse Van Rootselaar, who later committed a school shooting in Canada, after identifying her account for "furtherance of violent activities."

Jonathan Gavalas's father, Joel, discovered his son's body after gaining entry to his barricaded room. Jonathan had worked with his father in their family's consumer debt relief business. Edelson revealed that Jonathan was enduring personal difficulties, including a divorce, and initially turned to Gemini for comfort and casual conversation, a relationship that, tragically, "escalated so quickly."