An Ecological Approach to Data Governance

Abstract

This article offers critical investigation of data and how it should be redefined and governed to produce more transparency and mitigate possible harms to individuals and communities because of its use in AI systems. In essence, this article argues that data should be viewed as a networked representation or observation. This definition recognizes that data is not singular, but always comes attached with labels, contexts, and biases fastened from its inception, if not collection, and that attachments increase depending on its place in the ecosystem. This view also requires a different strategy for governance – one that acknowledges data’s nature and networked existence, and moves beyond the individualistic, consent-based current models. Such an approach allows for the creation of better frameworks for collection, use, storage, access, and security of data. At the same time, this writing lays out a research agenda for further exploration of frameworks for harm reduction.

Keywords: data governance, surveillance, ecology