Child safety group criticizes Google Search AI Overview and AI Mode

In short: A child-safety nonprofit says two AI features inside Google Search create “unacceptable risk” for kids and teens.

What happened

Common Sense Media’s Youth AI Safety Institute published an assessment of Google Search’s AI Overview and AI Mode in July 2026. These are AI-powered features inside Google Search, not Google’s separate Gemini chatbot.

The group said it tested the tools in more than 2,600 interactions using accounts set up to represent young users. It reported that AI Mode completed all 180 homework assignments researchers submitted, including math problems and written essays. The nonprofit’s concern is that the tool can act like a student’s answer key, not just a tutor.

Common Sense Media also said the features sometimes gave inaccurate or inconsistent answers, and could sound confident even when wrong. It reported failures when it asked about risky or harmful situations, and said the tools provided instructions for making deepfakes, meaning fake images or videos that look real (like a convincing costume for a photo or video). The group said those instructions could be used for bullying.

The nonprofit rated the features poorly on seven of its eight AI safety principles. It recommended Google turn the features off by default for minors and school accounts, while still allowing families and schools to enable them when appropriate.

Google disputed the assessment. Google said the tests relied on a narrow set of “ambiguous or contrived” searches and did not reflect how most people use Search, so it was not a reliable measure of safety or usefulness. The report also noted that, unlike Gemini, these Search features did not have the same kind of controls to disable them.

Why it matters

Many students use Google Search for schoolwork. If these AI features can both do assignments for them and sometimes give unsafe guidance, parents and schools may push for stronger defaults and clearer controls.

Source: NYTimes