The Gemini AI allegedly planned the suicide of a 13-year-old girl

The parents of a 13-year-old American schoolgirl have filed a lawsuit against Google, accusing its artificial intelligence (AI) assistant Gemini, accessible on the computer provided by the school, of having helped her for months to plan a suicide and her funeral.

The complaint, filed Monday in a California court, is the first public case accusing Google of having “hastily” integrated its AI into Chromebooks, these low-cost computers distributed to millions of American students and the rest of the world.

According to the complaint, the establishment of the teenager, named Emma, ​​had blocked Gemini and ChatGPT, but could not disable the search engine’s AI. This will still be the case despite a turnaround announced Tuesday by Google: the group will now deactivate its AI tools for students by default, but schools will not be able to offer its search engine without its AI functions.

“Schools should turn off search and leave it disabled until Google removes AI mode from search, or ensures that every student is always safe when interacting with AI,” Andrew Liddell, one of the family’s lawyers, responded to AFP.

In the fall of 2025, the North Carolina middle school student was “a happy 13-year-old teenager,” a musician and a Girl Scout, “with no history of mental illness,” according to the complaint. In December, she began “a silent spiral towards a serious psychological crisis”, until exchanging around 3,200 messages with the AI ​​in February, often at night, according to her lawyers.

Gemini “diagnosed Emma with serious mental illnesses and persuaded her to trust him rather than her parents and school counselors,” the complaint alleges. On March 17, the school’s surveillance software spotted a document titled “Funeral” and the school alerted his parents. Her mother logged into her daughter’s account and watched her write a goodbye letter live, before her father intervened.

“We are reviewing the accusations contained in the complaint and we take them very seriously,” Google spokesperson José Castañeda told AFP. Google argues that AI mode has safeguards against self-harm and circumvention attempts. But Emma “was able to come back to it again and again,” replies Andrew Liddell. “Whatever the safeguards were, they were clearly insufficient.”

AI creators face a series of complaints in the United States. The parents of Adam Raine, a 16-year-old Californian who committed suicide in April 2025, accuse ChatGPT, the OpenAI assistant, of having accompanied their son in his action. Google and the start-up Character.AI agreed in January to settle amicably several complaints from families of adolescents.