Seminar

Abstract:

[nur in Englisch verfügbar]

The combination of increased availability of a large amount of fine-grained human behavioral data and advances in machine learning leads to a growing dependence on algorithms to solve complex societal problems. Algorithmic decision-making processes could lead to more objective and thus possibly fairer decisions than those made by humans, who may be influenced by greed, prejudice, fatigue or hunger. However, algorithmic decision-making has been criticized for its potential to increase discrimination, information and power asymmetry, and opacity. Furthermore, it is not clear what fairness actually is and it has been shown that some of the most obvious criteria for fairness cannot be met simultaneously.

In this talk I provide an overview of the available technical solutions to improve fairness, accountability and transparency in algorithmic decision making. I also emphasize the criticality and urgency of engaging multidisciplinary teams of researchers, practitioners, policy makers and citizens to jointly develop, deploy and evaluate algorithmic decision-making processes in the real world, aiming for maximum fairness and transparency.