Algorithmic Decision-Making, Ambiguity Tolerance and the Question of Meaning

Lisa Herzog*

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

11 Downloads (Pure)

Abstract

In more and more contexts, human decision-making is replaced by algorithmic decision-making. While promising to deliver efficient and objective decisions, algorithmic decision systems have specific weaknesses, some of which are particularly dangerous if data are collected and processed by profit-oriented companies. In this paper, I focus on two problems that are at the root of the logic of algorithmic decision-making: (1) (in)tolerance for ambiguity, and (2) instantiations of Campbell’s law, i. e. of indicators that are used for “social decision-making” being subject to “corruption pressures” and tending to “distort and corrupt” the underlying social processes. As a result, algorithmic decision-making can risk missing the point of the social practice in question. These problems are intertwined with problems of structural injustice; hence, if algorithms are to deliver on their promises of efficiency and objectivity, accountability and critical scrutiny are needed.

Original languageGerman
Pages (from-to)197–213
Number of pages17
JournalDeutsche Zeitschrift fur Philosophie
Volume69
Issue number2
DOIs
Publication statusPublished - 18-Apr-2021

Cite this