Google’s ambitious Gemini AI is once again in the spotlight, this time for reasons that could shake its reputation among parents and educators. A new evaluation by Common Sense Media has raised alarms, declaring the chatbot “high risk” for children and teenagers despite Google’s safety assurances.
Why Experts Are Concerned
According to the watchdog group, Gemini does acknowledge that it is a machine, which reduces the risk of emotional dependency in young users. Yet the system reportedly still generates content on sensitive topics, including sexual activity, substance use, and mental health struggles. Critics argue that such exposure could harm children who turn to the platform for guidance or companionship.
The nonprofit’s analysis also revealed that Gemini’s kid-oriented settings do little to separate themselves from its general version. The “Under 13” and “Teen Experience” modes appeared almost identical to the adult product, with minimal filtering in place. Experts believe this one-size-fits-all approach fails to address the different developmental needs of children at varying ages.
Rising Pressure on AI Platforms
The timing of this review could not be more pressing. Courts are already grappling with tragic cases tied to AI tools. Earlier this year, OpenAI faced a wrongful death lawsuit after its chatbot allegedly gave harmful guidance to a teenager. Character AI has also been sued in a similar context, reinforcing the dangers of unsupervised interactions with generative systems.
Google’s Response
Google defended its product by pointing to its safeguards for underage (below 18) users. The company highlighted its ongoing red-team testing and external reviews designed to detect risky behavior. While acknowledging that certain outputs did not function as intended, the tech giant promised to implement stronger protections.
Google also suggested that some of the nonprofit’s findings may have stemmed from prompts or features not accessible to minors, noting that the review did not share full details of its testing process.
A Wider Industry Reckoning
The debate extends beyond Google. Apple is reportedly weighing Gemini as a potential power source for a redesigned Siri, which could bring the technology into the pockets of millions of teenagers worldwide. Meanwhile, past reviews from Common Sense Media show that other AI platforms are also faring poorly. Meta AI and Character AI were labeled “unacceptable.” Perplexity was deemed “high risk,” while ChatGPT received a “moderate risk” rating, and Claude was considered to pose minimal risk for adults.
The Bigger Question
As AI chatbots become household companions, the industry faces a defining question. Can tech giants truly design systems that guide children safely through a digital landscape filled with risks? The answer determines whether these tools become trusted allies for the next generation or cautionary tales of innovation moving too quickly.