Project AI Evidence (PAIE)

Project AI Evidence (PAIE) identifies, evaluates, and scales innovative applications of AI for social good and in the fight against poverty by bringing together AI adopters, tech companies, and researchers.

Artificial intelligence has the potential to transform our lives. It is a general purpose technology that spans sectors and is widely accessible—in reach of anyone with a smartphone—and is constantly improving.

But as the AI technical revolution unfolds and AI makes its way into many more social programs, there is very little rigorous evidence on where and how AI can have an impact on social outcomes. Without such evidence, companies, social innovators, and regulators lack the information to maximize the benefits of AI while protecting against its downsides. 

PAIE provides an unparalleled opportunity to rigorously study how these technologies can help—or harm—the well-being of people, particularly people who experience poverty. This is especially important given the speed, scale, and potential reach of many of the applications being rolled out. As AI begins to be employed across the social sector, randomized evaluations of AI applications will help us understand the most effective and least harmful approaches.

Through PAIE, researchers across J-PAL’s network work with nonprofit organizations worldwide. PAIE also collaborates with Schmidt Sciences to study the impacts of generative AI in the workplace, particularly in lower- and middle-income countries.

PAIE works with researchers, policy partners, and AI solution developers to:
• Determine key innovations where AI is most likely to have the maximum social benefit based on more than two decades of our research across multiple sectors.
• Identify and support innovative organizations to use AI to advance social good.
• Fund rigorous research to measure the impact of AI solutions in priority areas.
• Scale and apply socially-beneficial AI solutions based on learnings from the research.
• Create and widely share a practical guide on “evaluating AI” to inform the wider research ecosystem on new methods and best practices for evaluating AI-based programs.
• Build capacity of researchers, governments, social enterprises, and NGOs to better incorporate and evaluate AI for social good.

The new initiative is prioritizing questions policymakers are already asking: Do AI-assisted teaching tools help all children learn? How can early-warning flood systems help people affected by natural disasters? Can machine learning algorithms help reduce deforestation in the Amazon? Can AI-powered chatbots help improve people’s health? In the coming years, PAIE will run a series of funding competitions to invite proposals for evaluations of AI tools that address questions like these, and many more.

While we are interested in AI's application in every sector, we anticipate that a lot of innovations will initially come from sectors that are seeing rapid AI adoption and/or likely to be most affected by AI, including education, health, labor markets, climate, and financial inclusion.