Study: Young users (9 to 18Y) ditch Google for AI, with unknown consequences

A systematic review of 173 studies on generative AI in education finds that AI is neither inherently beneficial nor harmful to students’ critical thinking and problem-solving skills - its effects depend on how it is used. The reviewers report a clear duality: AI supports thinking when it prompts questioning, comparison and construction of one’s own reasoning, but undermines independent problem solving when treated as a shortcut to ready-made answers. Quantitative summaries show mixed outcomes: of 139 studies on critical thinking, 67 reported only positive effects, 33 only negative, and 39 both; of 103 studies on problem solving, 57 were exclusively positive while 23 were negative and 23 reported mixed effects. Parallel behavioral shifts underline the stakes: Google search use among 9-18 year olds fell from 72% to 47% between 2024 and 2026, 39% of 11-12 year olds report using AI, and youth-facing sites lost traffic as users migrate to AI summaries.

The review exposes three urgent gaps: 80% of studies use university students rather than children and adolescents, most rely on self‑reports instead of longitudinal or objective measures, and research is geographically skewed toward parts of Asia with Africa, South America and Central Asia underrepresented. Practical implications are firm: teacher education must train educators to use and teach AI critically, design tasks that preserve essential cognitive processes, and restrict unguided AI use for very young pupils. Without intentional policy and pedagogy to cultivate AI literacy, uncritical adoption will deepen disparities and let technology shape learning rather than serve it.