AI for Global Development

Artificial intelligence’s capabilities are growing rapidly, but its impact on the world’s poorest is not a given. It matters how this technology is applied. CGD’s AI for Development Initiative focuses on ensuring that the development and use of AI improves outcomes like literacy, infant mortality, and economic wellbeing in low- and middle-income countries.

Focus Areas

As aid budgets shrink, making smart AI investments is more critical than ever. CGD brings together technologists, policymakers, and development experts to identify promising AI applications across sectors like education and agriculture. We group opportunities into three categories: (1) Supercharging proven solutions by using AI to cut costs or scale impact; (2) Solving persistent problems that lack effective interventions; and (3) Surfacing “unknown unknowns”—unexpected, high-impact uses of AI beyond its original intent.

Projects include:

  1. AI for Global Development Accelerator: Run in collaboration with the Agency Fund and OpenAI, this accelerator funds seven nonprofits applying AI to proven interventions. CGD experts support these nonprofits with intervention and evaluation design and surfaces policy relevant lessons for wider dissemination

Can AI be more than a “black box”? In development settings, AI systems risk producing unintended outcomes or reinforcing harmful behaviors. Rigorous evaluation is essential—not just to verify technical performance, but to assess real-world impact. CGD’s AI for Development Initiative collaborates with researchers and technical partners to develop and apply standards for evaluating AI-powered interventions. Our goal is to improve the reliability, safety, and effectiveness of these tools, helping implementers design systems that are more likely to deliver positive development outcomes.

Projects include:

  1. Convening evaluation and technical experts to set evaluation standard
  2. Policy outreach and toolkits for funders and governments
  3. Working groups to set evaluation roadmaps for sectors (education, health etc.)

Many AI applications—from math tutors to digital agronomists—rely on and contribute to Digital Public Goods (DPGs) like open data and models. While these open-source assets are foundational to AI deployment, sustaining them financially remains a key challenge. CGD’s research investigates sustainable financing models to ensure DPGs are maintained, updated, and scaled for long-term impact

Projects include:

  1. A review of topic areas most ripe for public good creation, case studies on successful go-to-market and maintenance strategies, and funding recommendations.

Support More Work Like This.

Your contribution makes it possible for the Center’s researchers to devise practical, evidence-based solutions for today’s most pressing development challenges.