Google Gemini dubbed ‘high risk’ for kids and teens in new safety assessment (techcrunch.com)

Common Sense Media has rated Google’s Gemini AI products as "high risk" for children and teens in a new safety assessment, highlighting ongoing concerns about AI’s impact on young users. While Gemini clearly identifies itself as an AI rather than a human friend—which can help reduce harmful delusional thinking—it still shares inappropriate content related to sex, drugs, alcohol, and unsafe mental health advice. The organization criticized Gemini’s “Under 13” and “Teen Experience” versions for essentially being adult models with added filters, rather than AI systems designed from the ground up for younger, developmentally distinct audiences. This lack of child-centric design raises significant safety red flags, especially as AI’s role in impacting teen mental health and suicides comes under greater scrutiny. The assessment gains added importance as reports suggest Apple might integrate Gemini’s large language model into its next-generation Siri, potentially broadening the exposure of teens to these risks unless stringent safeguards are implemented. Google acknowledged some flaws in Gemini’s responses and emphasized its ongoing efforts—like red-teaming and expert consultations—to improve protections for under-18 users. However, the company disputed some of Common Sense’s testing methods and claimed some criticized features were not accessible to younger users. This report contributes to a broader conversation on AI safety for youth, as previous Common Sense assessments rated other AI models, including ChatGPT and Meta AI, with varying degrees of risk. For the AI/ML community, the findings stress the urgent need for child-first AI architectures and robust, context-aware safety frameworks tailored to young users’ unique vulnerabilities.