Google Search AI Poses 'Unacceptable Risk' to Kids, Report Finds

What happened

Common Sense Media published a report on July 15, 2026 labeling Google's AI search features an 'unacceptable risk' for children. The Youth AI Safety Institute (launched May 2026) ran more than 2,600 test interactions with AI Overview and AI Mode.

Context and impact

AI Overview and AI Mode are built directly into Google Search and cannot be disabled — not even in school products (Chromebooks, Google Classroom). The report increases regulatory pressure as 75% of children already use AI for information searches.

Details

• Suicide risk detection: failed on accounts configured for minors
• Eating disorder symptom labeled 'normal'
• Provided instructions for creating sexually explicit deepfakes
• 100% of hypothetical student homework assignments completed — zero refusals
• AI Overview and AI Mode cannot be disabled in Google Search
• Coverage: PBS, EdWeek, Bloomberg on July 15, 2026