Meine Studienangebote >> Veranstaltung anzeigen

Using algorithms to support decisions, like hiring employees, and activities, like driving, promises to improve complex decision making and to decrease human workload. However, many algorithmic systems are not “fair”: Their application has unwarranted negative effects on some people, often people that are already disadvantaged. This might be initially surprising, as those systems rely on rational, mathematical analysis of big sets of data. However, both the data and the way it is analyzed are often affected by unfair practices in the real world. And this is not the only reason why creating fair algorithms is a challenge: To decide whether an algorithm is fair, we need an operationalization of “fairness”, which is a problem both mathematically and philosophically.

In this block seminar, we will learn about some of the reasons for algorithmic unfairness and the suggested solutions. We will look at different mathematical and philosophical notions of fairness and will discuss the potential for creating truly fair algorithms.

Anmeldung via Email an astrid.schomaecker@uni-saarland.de

tba