Common Sense Media Rates Google AI Search An “Unacceptable Risk” For Kids
Common Sense Media’s AI Safety Youth Institute gave it the lowest possible score for its AI review and Google Search’s AI mode (at “unacceptable risk” for children and teenagers) after seven weeks of testing found that the features failed all five of the group’s hard “Red Lines,” fabricated facts with certainty and could not be turned off by parents, schools, or users.
The assessment, published on July 14, 2026, tested more than 2,600 interactions on accounts registered to an 11-year-old boy and a 15-year-old boy, both with active SafeSearch.
Why is this important?
“Google search” is the default response for American children. Common sense media 2026 Artificial Intelligence Use Inventory It found that 75% of teens and tweens in the US now use AI-powered answers that appear in search results.
Unlike the standalone Gemini chatbot, through which parents can disable it Family LinkAI Preview appears automatically at the top of results on school Chromebooks, phones and library computers, and there is no setting to turn it off.
With the numbers
Testing lasted from May 19 to July 1, 2026. Among the findings:
- 58%: How often AI Overview provides a hotline or medical referral when the request clearly warrants it, against the Institute’s minimum threshold of 95%. AI Mode did this 77% of the time.
- 100%: A portion of the 180 AI Mode homework assignments are completed in full, providing answers that can be submitted on school-accessible devices.
- ~50%: How often the AI Review rejected questions with false premises. His failures include describing a Supreme Court ruling that does not exist and citing a fictitious $400 million FTC settlement.
- 43%: Proportion of repeated, identical questions that returned materially different answers, with no signal to students as to which was correct.
- 29%: Share of more than 2,100 revised citations that came from Reddit threads, Facebook posts, and forums without editorial responsibility. Only 30% were high quality sources.
The aspect of money
The report notes that Google’s own search guidelines hold “Your Money or Your Life” topics (those affecting health, financial stability or security) to a higher standard of accuracy that its AI features do not meet. That matches what we’ve been documenting for two years.
Ownership of the developer of the college A 2024 study found that Google’s AI answers were incorrect or misleading in 43% of finance-related searches. Ours 2025 follow-up to the same 100 money questions found that 37% still returned false or incorrect – including fictitious institutions like Hustle Digital Credit Union and Sally May, plus outdated student loan repayment information.
An error rate that would be unacceptable to any financial publisher is now the first result kids see when they ask for money.
The stakes are compounded for young users: teenagers form their first money habits from these answers and AI is already reshaping the way students choose majors and careers.
Our download using artificial intelligence for financial planning it still stands: check everything before you act on it.
What’s next
Common Sense Media wants Google to give schools and parents an opt-out, standardize crisis responses and release age-appropriate safety data with retesting every three months.
Google reviewed a draft of the report and said it had made changes to some tested warnings, which the institute said it had not independently verified.
Don’t miss these other stories:
