News

By Gabriele Caldas Cabral and Steve Powell, Causal Map

How can we measure the impact of development programmes without losing the voices and experiences of the people they’re designed to support?

As conversations around the ethical implications of AI continue across the global justice space, its potential role in monitoring and evaluation raises both opportunities and important questions. In a new research paper, SWIDN members Gabriele Caldas Cabral and Steve Powell explore whether AI could help make evaluation more inclusive, while challenging traditional approaches to measuring impact.

Too often, development programmes are evaluated against predefined targets and indicators, shaped by the priorities of funders rather than the experiences of local communities. This can mean important stories, unexpected outcomes and local perspectives are overlooked.

Their paper, Listening at Scale: AI-assisted Causal Mapping for Inclusive Evaluation, explores a different approach. By using AI to identify connections and patterns across large volumes of interviews and personal accounts, researchers can better understand how people themselves describe the changes happening in their lives, without forcing their experiences into predetermined categories.

Drawing on a case study of SWIDN Member, Tree Aid’s work in Burkina Faso, the authors demonstrate how this approach can uncover outcomes that conventional evaluation methods might miss.

Importantly, the paper also recognises the limitations and biases of AI. Rather than replacing human judgement, the authors argue that AI, used carefully and with appropriate oversight, could help amplify perspectives that have historically been marginalised in development research.

The research raises important questions for the international development sector about whose knowledge counts, who gets to define success, and how we can make evaluation more accountable to the communities at the heart of development work.

Read the full paper: Listening at Scale: AI-assisted Causal Mapping for Inclusive Evaluation

Photo Credit: Claudio Schwarz