A report released by Common Sense Media on July 15, 2026, alleges that Google’s AI search features, AI Overview and AI Mode, routinely failed to identify harmful behaviors in test interactions. According to the study, which utilized 2,600 test queries from accounts designated as 11 and 15-year-olds, the tools failed to recognize 29% of explicit references to suicide and half of indirect references. The report stated that the AI features also provided full answers for hypothetical homework assignments and gave inconsistent responses to questions regarding mental health.
Google defended the technology in a statement, asserting that its AI features provide "extra layers of protection" beyond standard search guardrails. The company characterized the Common Sense Media report as using "ambiguous and contrived queries" that do not accurately represent real-world use. Google further noted that parents maintain controls to disable Search, though the report noted that the specific AI summaries within the search interface cannot be individually turned off by users.
The findings come as U.S. lawmakers consider new regulations for AI literacy and safety. According to a previous Common Sense Media survey, approximately 75% of American children aged 9 to 17 use AI-generated search summaries. Education experts noted that Google’s presence in thousands of schools makes the accuracy of these default tools a significant concern for student safety and academic integrity.