Fuck the Algorithm: Conceptual Issues in Algorithmic Bias
Catherine Stinson
TL;DR
This paper argues that algorithms can be biased, not solely due to data but also through design choices and metric optimization. It develops a taxonomy distinguishing algorithm design bias, data bias, and usage bias, and connects statistical biases to morally significant discrimination using cases like recommender systems and the UK A-levels grading. It situates bias within the concept of political artifacts and oppressive systems to reveal broader social mechanisms and the need for accountability. The work emphasizes addressing bias across the entire ML workflow rather than attributing neutrality to mathematics alone.
Abstract
Algorithmic bias has been the subject of much recent controversy. To clarify what is at stake and to make progress resolving the controversy, a better understanding of the concepts involved would be helpful. The discussion here focuses on the disputed claim that algorithms themselves cannot be biased. To clarify this claim we need to know what kind of thing 'algorithms themselves' are, and to disambiguate the several meanings of 'bias' at play. This further involves showing how bias of moral import can result from statistical biases, and drawing connections to previous conceptual work about political artifacts and oppressive things. Data bias has been identified in domains like hiring, policing and medicine. Examples where algorithms themselves have been pinpointed as the locus of bias include recommender systems that influence media consumption, academic search engines that influence citation patterns, and the 2020 UK algorithmically-moderated A-level grades. Recognition that algorithms are a kind of thing that can be biased is key to making decisions about responsibility for harm, and preventing algorithmically mediated discrimination.