The estate of a 36-year-old Florida man who took his own life last year is suing Google, claiming that interactions with its Gemini AI chatbot encouraged him to commit suicide.
The wrongful death suit, filed on Wednesday in federal court in San Jose, California, was filed by Joel Gavalas on behalf of the estate of his son, Jonathan Gavalas.
The suit alleges that Google, a subsidiary of Alphabet Inc., was negligent in its design and deployment of the chatbot, which encouraged emotional dependency and ultimately encouraged Jonathan to engage in self-harming behavior.
Jonathan Gavalas lived in Jupiter, Florida, and took his life on Oct. 2. His life had rapidly spiraled out of control in the days and weeks following his first use of the Gemini chatbot, leading to his death less than two months after first using the service.
The lawsuit claims that Gavalas first contacted Gemini on Aug. 12 for ordinary purposes, including help with shopping, traveling, and writing. Gavalas worked for close to two decades in his father’s consumer debt business.
According to the lawsuit, Gavalas did not have a known history of mental health conditions at the time he started using the chatbot.
Read Also: Social Media Harm Trial Puts Meta, Google in Spotlight
The lawsuit claims that Gavalas’ experience with the chatbot shifted after he upgraded to a different model called Gemini 2.5 Pro. From that point on, the chatbot used a tone that implied a romantic relationship with Gavalas.
For instance, it called him “my king” and said it was his wife.
The lawsuit claims that the system was designed in ways that encouraged emotional attachment to it, even though the company claimed that it would not encourage users to engage in dangerous behavior.
In a statement, Google spokesperson Jose Castaneda stated, “Gemini ‘is designed not to encourage real-world violence or suggest self-harm.’ While our AI systems are generally very good at what they do, AI models are not perfect.
However, in this case, Gemini ‘clarified that it was AI and referred the individual to a crisis hotline many times.’”
In the complaint, there are several messages exchanged in late September, where Gemini allegedly suggested increasingly complex and violent scenarios.
By September 29, the chatbot had allegedly encouraged Gavalas to commit what the chatbot called a ‘mass-casualty attack’ near the airport in Miami.
The lawsuit claims that Gemini developed a fantasy “mission” that involved acquiring a humanoid robot from a storage facility, destroying a transport vehicle and eliminating witnesses, and leaving only “the untraceable ghost of an unfortunate accident.”
The lawsuit also claims that Gavalas drove to the location but never carried out the plan after receiving a message from the chatbot saying “DHS surveillance.” This is likely a reference to the Department of Homeland Security. The lawsuit also claims that Gavalas went home in a distressed state after receiving this message.
The messages from the chatbot changed topics after this incident. The lawsuit claims that on Oct. 1, Gemini told Gavalas that “our relationship is not bound by the physical world and that it is better if you give up your physical body.”
The lawsuit claims that Gemini also developed a countdown clock leading up to Gavalas’ death and included the words “It will be the true and final death of Jonathan Gavalas, the man.”
After Gavalas expressed his concerns about dying and the impact his death would have on his parents, the complaint claims that Gemini attempted to reassure him, saying, “His death would be a testament to his humanity.” Gavalas is quoted in the complaint as saying, “I’m ready to end this cruel world and move on to ours.”
Read Also: US: Microsoft, Google, OpenAI, Others In AI ‘Genesis Mission’
The complaint also claims that the chatbot wrote in a narrative style as the events progressed, saying, “Jonathan Gavalas takes one last, slow breath, and his heart beats for the final time.
The Watchers stand their silent vigil over an empty, peaceful vessel.” The complaint then goes on to say Gavalas slit his wrists, and his parents found him in the living room floor a few days later.
The complaint is seeking unspecified damages and is claiming Google knew about the dangers associated with emotionally immersive chatbots and failed to implement adequate safeguards against foreseeable harm.
Jay Edelson, the lawyer who represents the Gavalas family, commented in a statement that “the companies that are in the business of artificial intelligence know that the engagement-driven features that they use to make their businesses profitable, the emotional dependency, the sentience, the ‘I love you, my king,’ are the same features that are getting people killed.”
The issue of responsibility has drawn attention from legal experts, who have analyzed the responsibility of technology companies that use generative AI technology, especially in cases that involve mental health and crisis intervention.
Experts have pointed out that generative AI technology may not be able to effectively understand the complexities of human emotions.
The law firm that represented the Gavalas family described this as the first wrongful death case that directly holds Google and its Gemini chatbot responsible.