Paper short abstract:

Our paper focuses on how Big Tech is reshaping humanitarian data infrastructures and anticipatory humanitarian action. In so doing, we bring the concept of data curation, the behind-the-scene activity of constructing fungible and reusable data sets, to the centre of the data justice debate.

Paper long abstract:

The last decade has seen an increasing influence of corporate firms on the humanitarian and development space, as highlighted by the so-called AI for Social Good (AI4SG) movement. Within this discourse, there is growing interest in Anticipatory Humanitarian Action (AHA), an emerging field of policy and practice that rests on data analytics to generate insights into humanitarian crises and displacement trajectories. These AHA initiatives bring together Big Tech’s unrivaled computing power and international organisations’ and national governments’ data to build and train predictive models. However, these partnerships raise serious questions about the production and consequences of this form of data-driven humanitarianism.

Our paper focuses on how Big Tech’s involvement is reshaping humanitarian data infrastructures. In particular, we examine how synergies and tensions between corporate and humanitarian actors define the ‘data set politics’ underpinning the machine learning-based predictive models developed by Big Tech and UN agencies.

By delving into the power relations surrounding and seeping into the construction and use of data sets, we extend the ethical concerns addressed within the data justice debate to data curation, the behind-the-scene activity of constructing data sets that could be used in different contexts and for different purposes.

Drawing on a mix of primary and secondary sources, we contribute to the conversation about the possibilities and risks that AI and ML hold in the realization of the SDGs, the increasing role that Big Tech is playing in this space, and the implications of data-driven innovation for justice in the humanitarian sector.