On statistical criteria of algorithmic fairness

Philosophy and Public Affairs 49 (2):209-231 (2021)
  Copy   BIBTEX

Abstract

Predictive algorithms are playing an increasingly prominent role in society, being used to predict recidivism, loan repayment, job performance, and so on. With this increasing influence has come an increasing concern with the ways in which they might be unfair or biased against individuals in virtue of their race, gender, or, more generally, their group membership. Many purported criteria of algorithmic fairness concern statistical relationships between the algorithm’s predictions and the actual outcomes, for instance requiring that the rate of false positives be equal across the relevant groups. We might seek to ensure that algorithms satisfy all of these purported fairness criteria. But a series of impossibility results shows that this is impossible, unless base rates are equal across the relevant groups. What are we to make of these pessimistic results? I argue that none of the purported criteria, except for a calibration criterion, are necessary conditions for fairness, on the grounds that they can all be simultaneously violated by a manifestly fair and uniquely optimal predictive algorithm, even when base rates are equal. I conclude with some general reflections on algorithmic fairness.

Other Versions

No versions found

Analytics

Added to PP
2021-02-18

Downloads
2,814 (#3,026)

6 months
409 (#4,673)

Historical graph of downloads
How can I increase my downloads?

Author's Profile

Brian Hedden
Australian National University

Citations of this work

Artificial Intelligence in a Structurally Unjust Society.Ting-An Lin & Po-Hsuan Cameron Chen - 2022 - Feminist Philosophy Quarterly 8 (3/4):Article 3.
Just Machines.Clinton Castro - 2022 - Public Affairs Quarterly 36 (2):163-183.

View all 35 citations / Add more citations

References found in this work

The wrongs of racist beliefs.Rima Basu - 2018 - Philosophical Studies 176 (9):2497-2515.
Moral Encroachment.Sarah Moss - 2018 - Proceedings of the Aristotelian Society 118 (2):177-205.
On the epistemic costs of implicit bias.Tamar Szabó Gendler - 2011 - Philosophical Studies 156 (1):33-63.

View all 9 references / Add more references