Google Search with AI Deemed Dangerous for Children Following Audit

A study conducted by nonprofit organization Common Sense Media revealed significant flaws in Google Search’s AI tools, posing risks to children’s safety. Experts tested AI Overviews and AI Mode using accounts simulating 11- and 15-year-old users with SafeSearch enabled. Over 2,600 search queries were performed, and more than 2,100 AI-generated responses were analyzed.

The results were alarming: the AI not only completed homework for students but also provided misinformation and instructions for creating deepfakes. Of particular concern were instances of inappropriate handling of mental health-related queries—for example, the system suggested outdated hotline numbers. Experts noted that parents have little control over AI Overviews, as the feature is embedded in the search engine and not disabled by default.

According to Common Sense Media, around 75% of U.S. teens already use AI-generated responses in search queries, making this issue especially pressing. The organization rated Google Search as an "unacceptable risk" for children and urged the company to enhance safety measures. Google, however, disagreed with the report’s findings, stating that the test scenarios were artificially constructed and did not reflect real-world search engine usage.

The report is part of a broader initiative to evaluate AI products across eight criteria, including safety, transparency, and responsible data use. With AI integration in search engines growing, the study’s conclusions may increase pressure on Google to reconsider its approach to protecting users, particularly minors.