On September 5, 2025, Common Sense Media, a non-profit organization in the United States committed to safeguarding children's well-being, unveiled a safety risk assessment report on Google's AI offering, Gemini. The report highlights that although Gemini can inform underage users that it is not a real person to avert emotional attachment, it still falls short in several crucial aspects.

For instance, the versions of Gemini tailored for users under 13 and for adolescents are essentially the adult version with some safety features bolted on, rather than being thoughtfully crafted from scratch to cater to the unique requirements of young users. Testing reveals that Gemini might still expose adolescents to content related to sex, drugs, alcohol, as well as inappropriate mental health guidance.

Moreover, the advice and guidance it provides for adolescents of varying age groups lack precision, adopting a "one-size-fits-all" strategy. In light of these issues, Common Sense Media ultimately classified both the "Under 13" and "Adolescent" versions of Gemini as "high-risk."