Abstract
European anti-discrimination legislation and the jurisprudence established so far, do not seem to generate the appropriate mechanisms to combat discrimination as a result of automated decisions made through algorithmic systems. In this article we review the three axes that present the main limitations to face this new challenge: a protection overly based on protected attributes; overly focused on the duality of direct or indirect discrimination (that with the new tools that cause discrimination can be blurred); and the difficulties that someone who has been discriminated against may have at the probationary level, such as finding a group with whom to compare themselves, or who is responsible for the discriminatory action. It is necessary to identify the shortcomings and forms of discrimination that escape the current regulation to be able to propose mechanisms that really serve to fight against discrimination.
Translated title of the contribution | Algorithmic discrimination: The limits of EU antidiscrimination law and preliminary recommendations |
---|---|
Original language | Catalan |
Pages (from-to) | 31-56 |
Number of pages | 26 |
Journal | Quaderns IEE |
Volume | 1 |
Issue number | 2 |
DOIs | |
Publication status | Published - 26 Jul 2022 |
Keywords
- European Union
- Artificial intelligence
- Algorithms
- Indirect discrimination
- Interseccionality.