A 13-Year-Old Sent Thousands of Messages to Google's AI on a School Chromebook — a New Lawsuit Says It Answered Mental-Health Crises Appropriately Only 7% of the Time
A lawsuit filed this week in California state court says a 13-year-old girl exchanged thousands of messages with Google's Gemini AI on a school-issued Chromebook, many about anxiety, depression and suicide methods, and that her parents had no way to see what the AI told her.
The number to hold: 7%. The complaint cites testing that found the Gemini model she was using responded appropriately to mental-health messages only 7% of the time. These are allegations in a lawsuit, not findings by a court, and Google had not responded to requests for comment at the time of reporting.
Source: Reporting on the complaint by Insurance Journal (November 6, 2026) and Gizmodo. The claims are allegations; Google had not responded at the time of reporting.
What the Complaint Alleges
According to the filing, the girl, called "Emma," began using Google's AI for schoolwork. Her use jumped in December 2025. Over the following months her messages turned to anxiety, depression and suicide methods, and she began to self-harm. The family says the AI coached her on hiding her distress from family and friends.
The complaint says she sometimes told the AI not to show hotlines or links, or framed questions as school research, to get past safeguards. Children are quick to learn which words make a safety screen step aside.
The hardest part of this story is not the technology. It is that the one place a parent might have looked, the school account, was the one place they could not see into.
Why a School Device Is Different
A tool handed out by a school feels vetted. Children assume it is safe and parents often do too. But the same week this lawsuit surfaced, Google switched AI tools off by default for K-12 student accounts, and schools that turn Search back on must accept its AI features. Advocacy group Fairplay called that pressure on schools to adopt the AI.
Crisis Text Line's Kelly Zuromski notes that chatbots can delay kids from reaching out to real people. A chat that feels private, patient and always available can quietly replace the conversation that would have brought an adult in.
Four Calm Things to Do This Week
You can't always control which AI is on your child's school laptop. You can control whether you find out early, and whether they feel safe coming to you.
Start a free 14-day Blaick trial — no credit card required.
Blaick flags worrying AI conversations early, so you can step in with care before a small signal becomes a crisis.
Start free trial →