Fuck the Algorithm: Conceptual Issues in Algorithmic Bias

TL;DR

This paper critically examines the conceptual underpinnings of algorithmic bias, clarifying misconceptions about whether algorithms themselves can be biased and exploring the moral implications of statistical biases.

Contribution

It provides a nuanced analysis of what algorithms are and how bias can be morally significant, connecting conceptual issues with real-world examples of bias in various domains.

Findings

Algorithms can be a locus of bias, affecting moral responsibility.

Bias in algorithms can stem from statistical data biases.

Understanding algorithms as artifacts helps address bias and discrimination.

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.