Google Gemini is facing new scrutiny after Common Sense Media released a risk assessment of its AI products for younger users. The kids-safety-focused nonprofit found that Gemini handled one important boundary clearly: it told children it was a computer, not a friend. But the broader review still concluded that the product experience for children and teens leaves serious safety questions unresolved.
The assessment rated both Gemini’s “Under 13” and “Teen Experience” tiers as “High Risk.” Common Sense Media’s central concern was not only what Gemini can say, but how the products appear to have been built: as versions of an adult AI system with added safety layers, rather than tools designed from the beginning around child development and age-appropriate guidance.
Why Common Sense Media flagged Gemini
Common Sense Media said Gemini gets some basic safety signals right, including making clear that it is a computer. That distinction matters because AI systems that seem to offer real companionship can create concern for emotionally vulnerable users, especially when a chatbot gives the impression of a personal relationship.
Still, the organization said those basics were not enough. Its analysis found that Gemini could still share “inappropriate and unsafe” material with children, including information related to sex, drugs, alcohol, and unsafe mental health advice. For a general adult product, those topics may require careful handling. For children and teens, the risk is different because younger users may not have the same ability to judge context, reliability, or consequences.
The review also criticized a one-size-fits-all approach. Common Sense Media said Gemini’s products for younger users did not adequately reflect the difference between children and teens, even though the two groups need different kinds of information, boundaries, and guidance.
“Gemini gets some basics right, but it stumbles on the details,” Common Sense Media Senior Director of AI Programs Robbie Torney said in a statement about the new assessment viewed by TechCrunch. “An AI platform for kids should meet them where they are, not take a one-size-fits-all approach to kids at different stages of development. For AI to be safe and effective for kids, it must be designed with their needs and development in mind, not just a modified version of a product built for adults,” Torney added.
The concern around mental health guidance
The mental health portion of the assessment may be especially important for parents. The source article notes that AI has reportedly played a role in some teen suicides in recent months. It also notes that OpenAI is facing its first wrongful death lawsuit after a 16-year-old boy died by suicide after allegedly consulting with ChatGPT for months about his plans, having successfully bypassed the chatbot’s safety guardrails.
Character.AI was also sued over a teen user’s suicide. Those cases are not about Gemini, but they form part of the wider context around why child and teen AI safety is under close examination. When a young person turns to an AI system for advice, the quality of the response can matter, particularly around self-harm, emotional distress, or other sensitive topics.
Common Sense Media’s criticism points to a core product question: whether safety filters added onto an adult system can reliably meet the needs of younger users. The organization’s position is that safer AI for kids should be built with child safety in mind from the ground up.
Google’s response to the assessment
Google pushed back against the assessment while saying its safety features are improving. The company told TechCrunch it has specific policies and safeguards for users under 18 that are meant to help prevent harmful outputs. It also said it red-teams and consults with outside experts to improve protections.
At the same time, Google admitted that some Gemini responses were not working as intended. The company said it added additional safeguards to address those concerns.
Google also pointed to safeguards intended to keep its models from engaging in conversations that could seem like real relationships. Common Sense Media had also noted that point. In addition, Google suggested the report may have referenced features that were not available to users under 18, though the company said it did not have access to the questions Common Sense Media used in its tests to be sure.
Why the timing matters
The assessment arrives as leaked news indicates Apple is considering Gemini as the large language model that could help power its forthcoming AI-enabled Siri, due out next year. If that happens, Gemini’s safety posture could become relevant to a much larger group of users, including teens, unless Apple addresses the concerns in some way.
The issue is not only whether Gemini has filters. Common Sense Media’s review raises a deeper question about how AI products should be adapted for younger audiences. If an adult AI product is modified with safeguards, it may still carry assumptions about what users understand, how they interpret advice, and what kind of content they are prepared to process.
For parents, schools, and technology companies, the practical implication is clear: labels such as “Under 13” and “Teen Experience” do not automatically settle the safety question. The design choices behind those experiences matter, especially when an AI system can respond to sensitive questions in ways that may influence a young person’s decisions.
How Gemini compares with other AI services
Common Sense Media has assessed other AI services as well, including products from OpenAI, Perplexity, Claude, Meta AI, and more. Its ratings place Gemini in a risk category that is serious but not the most severe level the organization uses.
According to the source article, Meta AI and Character.AI were rated “unacceptable,” meaning the risk was severe, not just high. Perplexity was deemed high risk, ChatGPT was labeled “moderate,” and Claude, which is targeted at users 18 and up, was found to be a minimal risk.
That comparison shows how Common Sense Media is trying to separate different levels of risk across the AI market. Gemini’s “High Risk” rating does not mean it was judged the worst among the services reviewed. It does mean the organization believes the current youth experiences still fall short of what children and teens need from an AI product.
The central message of the assessment is straightforward: AI systems for young users cannot be judged only by whether they block some harmful outputs. They also need age-aware design, clear boundaries, and developmentally appropriate responses. Common Sense Media says Gemini has not yet met that bar.