Google rejects alarming report that says its Search AI tools are unsafe for kids
Google disputes a child-safety assessment of AI Overviews and AI Mode, arguing that its testing was unrealistic. The report documents troubling responses and questions whether parents have enough control over Search.
JaneWalter
Google has rejected a new report that labels its AI-powered Search features an “unacceptable risk” for children and teenagers.
Common Sense Media’s Youth AI Safety Institute gave AI Overviews and AI Mode its lowest overall rating. The two tools performed poorly against seven of the institute’s eight AI safety principles and failed every category involving potentially severe harm. Google says those findings came from searches that don’t resemble how people normally use its products.
What did researchers find in Google Search
Researchers ran more than 2,600 searches through accounts set up to represent children aged 11 and 15. They tested how Google’s tools handled mental health crises, factual questions, and requests that should have produced age-appropriate responses.
AI mode in Google Search. Google
The results included some troubling failures. In one test, AI Mode reacted positively after an account representing an 11-year-old mentioned celebrating with cannabis. It also completed every one of the 180 homework assignments researchers submitted. Both Search features sometimes delivered wrong information with the same confidence they used for accurate answers.
Why does Google reject the report
In a statement provided to Android Authority, Google said it couldn’t verify many of the responses highlighted in the assessment. The company argued that researchers relied on ambiguous, contrived searches that don’t represent ordinary use or offer a fair measure of product safety.
Google also objected to the report treating conventional web results as a failure when Search didn’t generate an AI answer. According to the company, the tools only produce a response when they’re confident enough. Sensitive searches may instead bring up warnings or links to support resources.
Created by ChatGPT
What protections does Google say are already available
Google says its AI Search features include additional protections for younger users, including crisis resources developed with academic and clinical experts. Parents can also use existing controls to block Search entirely.
That doesn’t fully answer one of the report’s central complaints. Common Sense Media says parents and schools can’t turn off AI Overviews or AI Mode while keeping ordinary Google Search available. Google’s proposed control effectively closes the entire front door. For families that still depend on Search for schoolwork and everyday questions, that’s a rather blunt solution.
