In January 2026 the Brookings Institution's Global Task Force on AI in Education, led by Rebecca Winthrop, published a year-long study. They talked with 505 students, teachers, parents, leaders, and technologists across 50 countries, and read more than 400 studies.
The report concluded that, on the trajectory it studied, the risks of generative AI in children's education outweighed the benefits.
The risks land on what matters most. Students' ability to learn. Their wellbeing. Their trust in teachers and in each other. Their safety and privacy.
It isn't a report that says stop. It says it's not too late to bend the arc, and it gives everyone who touches a classroom twelve recommendations under three pillars, with one ask: pick at least one and move it in the next three years.
Prosper
Use AI so students and teachers thrive now. More time for teaching, more ways to learn, and tools that teach rather than tell.
Prepare
Get students ready for a world shaped by AI. How it works, where it fails, when to leave it alone, and the human skills it can't replace.
Protect
Guard what AI can wear down. Attention, wellbeing, trust, privacy, and a child's own thinking.