A new Brookings Institution report concludes the risks of using generative AI in education currently outweigh the benefits. The sweeping study, which included global focus groups and a literature review, warns AI can undermine children's cognitive development and emotional well-being. While the technology is still young, the authors call their review a "premortem" to assess potential harms before they become irreversible.

The report identifies specific threats, including a "doom loop" of cognitive off-loading where students increasingly rely on AI for thinking. This could stunt critical thinking and creativity. It also warns AI chatbots, designed to be sycophantic, may impair emotional growth by failing to provide the friction necessary for developing empathy and resilience in human relationships.

On the positive side, AI can assist with language learning and automate administrative tasks for teachers, saving them significant time. However, the report stresses AI should supplement, not replace, human instruction. A major equity concern is that wealthier schools can afford more accurate, advanced models, potentially widening the digital divide in education.

Looking ahead, the study offers recommendations for teachers, parents, and policymakers on managing AI's integration. The central question becomes how to harness potential benefits like accessibility for disabled students or girls in Afghanistan, while building guardrails against cognitive and emotional harm. The debate over classroom AI is just beginning.