Becoming Data Equals

Democratic Equality, Educational AI, and the Politics of Data Design

Abstract

In, Data Equals: Democratic Equality in Data Design, Colin Koopman investigates the impact and downstream consequences of artificial intelligence (AI) on data infrastructures in educational systems, democratic governance, and normative politics. Koopman’s main argument is that democratic equality must be actively built into data formats, classificatory systems and technical architectures from the outset, as opposed to treating them as neutral tools that often require downstream ethical correction. To illustrate this, Koopman focuses on educational AI and how personalized learning systems, that rely on “learner models” to render students computationally legible through selective features, simply manage to exacerbate existing inequalities and pedagogical assumptions. He then offers a “format anatomy” as a method to shift this critique upstream to the level of database schema and data ontologies to show how these structures privilege individualized rather than relational concepts of learning. This book review will discuss the distinction Koopman draws between technologies of separation and technologies of collaboration and how this offers a powerful framework for rethinking educational technology beyond current narrow concern such as bias, privacy, and efficacy. Although Koopman’s data design agenda remains predominantly conceptual, it does make a compelling contribution by advancing a democratic theory of data infrastructure.