Artificial Intelligence in Education Evidence Synthesis and Use in Low- and Middle-Income Countries

👉 Download the Executive Summary

👉 Download the Working Paper

Artificial intelligence is rapidly reshaping how education evidence is found, screened, synthesised, translated, and communicated. For education systems in low- and middle-income countries (LMICs), this shift presents a significant opportunity: AI tools can dramatically reduce the time and cost required to navigate large and fragmented evidence bases, making research more accessible to policymakers, practitioners, and funders. At the same time, the risks are substantial. Poorly governed AI use can introduce confident inaccuracies, obscure methodological weaknesses, and reinforce existing global inequities in whose evidence is visible and valued.

This working paper provides a practical, principles-based guide to using AI responsibly across the education evidence ecosystem in LMIC contexts. Drawing on a review of existing AI tools, emerging practice, and expert consultation, the paper:

• Maps current and emerging AI applications across the education evidence lifecycle
• Assesses risks, trade-offs, and limitations of AI-assisted evidence work
• Examines implications specifically for LMIC education systems and institutions
• Proposes a principles-based framework for responsible AI use

Rather than advocating for rapid or universal adoption, the paper emphasises fit-for-purpose use of AI, grounded in transparency, human oversight, and contextual awareness. It highlights the importance of safeguarding research quality, protecting local knowledge systems, and ensuring that AI strengthens—rather than bypasses—LMIC evidence institutions.

The paper concludes with a set of key questions for funders, researchers, intermediaries, and decision-makers seeking to harness AI to improve education outcomes while maintaining trust in evidence.