WP 1: 

The first work package addresses the fundamentals of protection against discrimination. The reason for this research is that decisions in typical cases of algorithmic discrimination have three characteristics:

Firstly, they are scalable and thus have an unprecedentedly broad impact; secondly, they are uniform and homogeneous; thirdly, they are inflexible and do not allow for deviations in individual cases.

This means that, in algorithmic decisions, the structural dimension takes on enormous significance alongside the risk of individual discrimination. Not only do material and compensatory disadvantages arise for individuals – allocation damages – but there are also immaterial and symbolic disadvantages for legally protected groups – representation damages. The latter, in particular, also pose a challenge to the conceptual design of anti-discrimination protection.

This research area builds on existing research into protection against discrimination. The aim is to identify which fundamental principles of non-discrimination law need to be adapted to the context of AI, and where translation is required.

To achieve this aim, we will cooperate with other subprojects (particularly SP 1). Furthermore, interdisciplinary exchange is crucial, particularly integration into the academic community for algorithmic fairness, so that legal scholarship can actively shape this discourse.