What turns development data into better decisions?

The panel connects development-finance data, synthesis of research and donor decisions. Yemesrach Workie describes IATI as an open standard and infrastructure whose usefulness depends on coverage, timeliness, completeness and supporting evidence. The financial totals in the transcript are unclear. Sophie Pascale presents Impact AI as a way to compare interventions across contexts; Kent Freeze, explicitly speaking personally, asks how investments connect to ultimate outcomes across sector boundaries.

Freeze argues for outcome evidence as well as input tracking, and warns that privileging RCTs can exclude qualitative and institutional questions that the design does not easily answer. This is a methodological scope argument, not rejection of trials. Pascale closes by distinguishing different kinds of failure.

Several distinctions recur. Institutional willingness to share data is not equivalent to capacity to organise it. Public availability does not establish useful coverage or transparency. Grassroots implementers can be missing because they lack publication resources, making visibility endogenous to institutional capacity. IATI's proposed quality classifications are future service improvements, not an already complete service. A network study is described as identifying connecting organisations, but network position alone does not demonstrate impact or justify allocation without further judgement.

The discussion then becomes political. Organisational disclosure rules shape what appears in common data. Freeze describes reputational and political risks to publishing failures, while Workie argues for greater transparency and civil-society pressure. This is a disagreement about how to respond to risk, not evidence that disclosure is costless. A question about stitching together health systems receives a candid response that common repositories may not always be the right approach and niche tools may remain useful. Another audience member challenges the assumption that faster answers to old questions adequately address AI's effect on the political information environment. Freeze acknowledges the gap between better evidence and actual decisions; evidence can reveal leaders' competing priorities rather than change them.

Funding comments are similarly qualified: philanthropic risk acceptance depends on the failure, organisational tolerance and the plausibility of the underlying theory. Claims that anticipation halves humanitarian expenditure or that a new philanthropic wave will resolve funding pressure are predictions or assertions, not findings demonstrated here.