Programs

Research and development areas

Four interconnected areas focused on understanding and improving AI's impact in education — from research to tools to practice.

01

Impact Research

Evidence on what actually works

We conduct and fund applied research on AI in education — efficacy studies, equity analyses, and measurement frameworks. Published openly for the entire field, not locked behind vendor partnerships.

  • —Efficacy studies of AI tutoring and assessment tools in real classroom settings
  • —Investigations of how large language models support (or hinder) deep learning
  • —Analysis of equity gaps in AI-powered education tools
  • —Development of measurement frameworks for AI learning impact
  • —Large-scale dataset analysis (16M+ student responses from Indian state assessments)

Evidence

AIED 2026 paper: 200 experimental conditions, 15+ models, LLM difficulty estimation. SmartPaper: measurable reduction in learning poverty across Rajasthan.

02

Research Tools

Open infrastructure for studying AI in education

We build and maintain open-source tools purpose-built for education research — anchored by Open Items, our open assessment platform built on the CZI Learning Commons Knowledge Graph.

  • —Open Items: 34K+ CC-licensed assessment items with AI generation, LLM evaluation, and adaptive practice (openitems.impact-edu.ai)
  • —Built on the Learning Commons Knowledge Graph — 250K standards, 2K learning components, 273K relationships
  • —LLM-as-Judge evaluation pipeline scoring on 5 dimensions with 85% auto-approve rate
  • —AI interview tools for qualitative research — text and voice
  • —Synthetic student simulation for instant psychometric feedback on new items
  • —UpGrade: open-source A/B testing for education (Gates Foundation + Schmidt Futures)

Evidence

Full K-12 curriculum content generation at $25-50 total cost. 98% mathematical accuracy. First applied project built on the Learning Commons Knowledge Graph.

03

Practitioner Training

Helping educators use AI with evidence, not fear

We train educators to be informed, critical, and effective users of AI — making evidence-based decisions about which tools work and which are hype.

  • —AI Literacy Lab for hands-on practice with conversational AI
  • —Professional development programs for teachers and instructional designers
  • —District-level AI integration planning support
  • —Train-the-trainer programs that scale through educator networks
  • —Practical guides and resources freely available online
04

Field-Building & Convening

Connecting the people who need to be in the same room

We strengthen the ecosystem connecting researchers, practitioners, developers, and policymakers — the people whose collaboration determines whether AI in education works for everyone.

  • —Annual convening of researchers and practitioners working on AI in education
  • —Working groups on critical topics: assessment, equity, safety, efficacy
  • —Policy briefs that translate research into actionable guidance
  • —Community of practice for educators implementing AI tools

Evidence

4 annual A/B Testing at Scale workshops at L@S (2020-2023). Deep relationships across CMU, TU Delft, BrainPOP, Carnegie Learning, Savvas, Pratham.

Want to partner on research?

We're looking for researchers, school districts, and funders.

Get in touch