AI Fairness - An Interdisciplinary Discussion

Fairness is a much-discussed concept in the current debate concerning AI Ethics. Many recent publications on AI fairness primarily focus on avoiding or fixing algorithmic-based discrimination of groups or individuals and on data-debiasing. These studies aim to offer various possible metrics, which can act as tools to evaluate whether groups or individuals are indeed treated differently. However, the multitude of existing indicators that label an AI system as “fair” or “unfair,” as well as the lack of standardized, application of field-specific criteria to choose among the various fairness-evaluation methods makes it potentially difficult – often extremely so – for AI auditors to judge whether or not certain algorithms result in discrimination or unfairness. Moreover, structural injustice and procedural fairness issues inherent to the software application environment should be considered, among other things, when evaluating the ethical impact of an AI system. 

This workshop aims to investigate the challenges represented by fairness evaluation of AI algorithms from an interdisciplinary perspective.