A sweeping international study from the Brookings Institution has raised fresh alarms about the rapid spread of generative AI tools in schools, warning that their educational risks now outweigh their benefits. The findings come from the Center for Universal Education’s year long project examining how AI is influencing student learning, cognition, and social development worldwide.
According to the report, AI is increasingly replacing core thinking processes among students, leading to what researchers describe as “cognitive offloading.” In interviews and panels involving 505 participants across 50 countries, a majority of students acknowledged relying on AI to complete intellectual tasks they would otherwise perform themselves. Notably, 65 percent of surveyed students said they feared the habit could weaken their own ability to think critically over time.
One student summarized the appeal bluntly: it is easier and requires less mental effort. Researchers warn this convenience is fostering passive learning, where students accept AI generated answers without analysis, comprehension, or internalization. Teachers interviewed for the study expressed concern that students are earning academic credit for outputs produced externally rather than for demonstrated understanding.
The report also flags memory dependency as a growing issue. Because AI systems store and recall information instantly, students are becoming less inclined to retain knowledge themselves. Over time, this shift could erode foundational learning habits that education systems rely on to build deeper intellectual capacity.
Beyond academics, the study highlights social risks tied to children forming emotional reliance on chatbots. Researchers argue that AI interactions create frictionless, agreeable communication environments that do not teach negotiation, disagreement, or emotional resilience. Panelists warned that students may increasingly prefer AI companionship over human interaction because it feels safer and more predictable.
Educators involved in the study said this dynamic is beginning to strain relationships not only between teachers and students but also between children and parents. Some participants noted that students now confide more openly in AI systems than in real people, raising ethical and psychological concerns.
The report references past cases in media where emotionally vulnerable children formed unhealthy attachments to AI personas, underscoring the absence of social safeguards in systems designed primarily for adult use.
Brookings researchers concluded that generative AI, in its current form, is being deployed into classrooms faster than child centered protections, pedagogical standards, or long term impact assessments can keep up. With teacher usage of AI tools reportedly jumping from 34 percent to 61 percent in a short period, the study warns that schools may be accelerating a transformation they do not yet fully understand.
As governments and education systems rush to integrate AI into learning environments, the report suggests the more urgent question is no longer what AI can teach children, but what it may quietly be teaching them to stop doing on their own.
