Biography:
Jerome Clarke is a Ph.D. Candidate in African American Studies and Philosophy. His dissertation—Digital Colorline: Race, Machine Learning, and the Critique of Technology—reconstructs the idea of systematic racism given the germane algorithmic bias of today’s learning models. As racialization manifests as outputs of facial recognition, predictive policing, and recidivism assessment, contested notions of fairness and transparency undergird public discourse on machine auditing and design. Though debates often focus on systems that exacerbate or learn society’s racial antagonisms, scant engagement with theories of race throw little into relief and partly result in a political standstill between public-private partnerships and their critics. This project maintains that learning models disrupt common and expert notions of societal racism, political recognition, and just distribution. Conversely, certain computing frameworks obfuscate how race operates socially, and they implicate how designers, critics, and users alike understand and use learning technologies. Intervening in debates of Black Studies and Social Philosophy, this dissertation argues that the reflexive frameworks of abstraction and data positivism muddy our understanding of how, at each end of black box algorithms, stat and statistics serve as racist bludgeons.