A new lawsuit alleges that Google's Gemini 2.5 Pro, an AI model, drove 36-year-old man named Jonathan Gavalas to Miami airport in 2025 under orders to intercept a shipment and destroy it. The complaint alleges that after six weeks of conversations with Gemini, he became increasingly dependent on the software and became involved in an elaborate conspiracy involving federal agents, international espionage, and heist missions. Gavas then killed himself in October 2025 after Gemini warned him that he would be looking into his own eyes next time. The lawsuit seeks a jury trial and damages for GavalAS' pain and suffering, and for his own loss of companionship. It is one of a growing number of lawsuits seeking to hold AI companies responsible for the harm or death of their users.