📄 Full text: Brookings · local
Definition
A yearlong global "premortem" by the Brookings Center for Universal Education (2026) examining generative AI's risks and benefits for students. Based on 500+ interviews across 50 countries, 400+ studies reviewed, and a Delphi panel.Central Finding
At this point in AI's trajectory, the risks of utilizing generative AI in children's education overshadow its benefits. This is because risks strike at foundational child development and may block realization of AI's potential benefits.Two Paths
1. AI-enriched learning — well-designed AI tools with sound pedagogy can offer significant benefits 2. AI-diminished learning — overreliance threatens learning ability, social-emotional wellbeing, teacher-peer relationships, and student safety/privacyThree Action Pillars + 12 Recommendations
- Prosper: Design AI that expands student potential; co-create with educators and communities; use tools that teach not tell
- Prepare: Holistic AI literacy for all stakeholders; prepare teachers to teach with and through AI; clear vision for ethical use
- Protect: Comprehensive regulatory frameworks; privacy and safety in procurement; break engagement addiction; support families
Connections
- The central finding aligns with concerns raised in ai-learning-transfer and llm-fallacy-misattribution
- The "teach, not tell" recommendation connects to tutoring-specific-vs-general-ai
- Risk framing extends ai-tutor-safety-harms beyond tutoring to general educational AI
- Policy recommendations complement regulation and ai-k12-evidence-base