The AI safety perimeter just moved into the one product kids cannot avoid

The kid-AI safety perimeter just moved from products children can avoid to a product they cannot. When the only search engine in a classroom is also an AI answer engine with no off switch, the rating stops being a product review and becomes a public-safety verdict. Mashable, reporting on Common Sense Media's Youth AI Safety Institute assessment, put Google AI Overview and AI Mode at the worst end of a five-step scale, not because the model is uniquely broken, but because the surface is uniquely unavoidable. A chatbot a child never opens can be rated and walked away from. A search box a child must open cannot. The mechanism is the wrapper: an opt-out becomes impossible when the AI lives inside a default utility. The constructive ask is a single, blockable switch for the AI components of Search. Without it, every classroom Chromebook ships with the same unrated chat attached to every query.