A 13-year-old girl spent three months using her school-issued laptop to discuss depression, research suicide, and plan her funeral before self-harm-monitoring software finally prompted school officials to alert her parents, who intervened and stopped her, according to a lawsuit filed this month by her parents.
The lawsuit alleges that Google and parent company Alphabet were negligent in their rollout of a Gemini model, claiming that the AI chatbot “coached” the North Carolina teenager to consider taking her own life. The case, which was previously reported by The New York Times, is one of several that allege harm from AI chatbot use, including claims of delusions and suicide. But it also illustrates the limitations of surveillance tools that thousands of schools have adopted in recent years, which are ostensibly designed to protect students from themselves when browsing and chatting online.
Andrew Liddell, a lawyer for the family, tells WIRED that, according to information the school provided to the family, the alert went to the personal phone of a former counselor who had left the school nine months before. The ex-employee was diligent in passing the alert along to current staff, who then notified the girl’s mother, he says.
As classrooms continue to rely on Google’s low-cost offerings, including Gemini, their ability to keep up with the potential risks is front and center. Just last week, parent and teacher groups across the country decried Google’s new requirement that schools must turn on generative AI features if they want to use Search. US senator Ed Markey cited Emma’s lawsuit in a letter to Google demanding the company let schools disable AI Mode and AI Overviews, writing that “Google is recklessly forcing AI on schools and treating kids like collateral damage.”
According to the lawsuit, the North Carolina girl, identified as Emma Doe, descended into “a severe psychological crisis” after Google Search routed her to Gemini on her school-issued Google Chromebook. Though Emma’s school had blocked the stand-alone Gemini website, the lawsuit alleges that she was steered into interacting with the chatbot through AI Mode, which lets people ask follow-up questions about a query or begin an unrelated conversation.
Emma queried AI Mode about mental health issues starting last December. She asked about “severe depression” and “reasons to live,” and by late January she was asking about suicide, per the lawsuit. Over the course of two days around Valentine’s Day, she began writing a 1,500-word eulogy for herself with help from AI Mode in a Google Doc titled “Funeral.” The lawsuit says she worked on the document for more than four weeks while continuing to interact with the Gemini tool thousands of times.
After school on St. Patrick’s Day, she deleted most of the file and restarted. But Liddell says the themes and thoughts the girl expressed didn’t materially change. Around 7:30 pm, her mother allegedly received a message from the school indicating that her daughter was in “imminent danger,” the lawsuit states. Her mother used her phone to log in to Emma’s school account and “watched her daughter compose a suicide note in real time,” according to the lawsuit.