22 April 2021
16:00 – 17:30 CET

Link to online seminars: Zoom

In less than two decades the output of machine learning algorithms has come to have a significant impact on human lives and society. In a wide range of domains such as health, criminal justice, and social services, the way in which ML algorithms rank, sort, filter, and categorize people form the basis of high-stakes decisions. It is widely appreciated that ML systems hold great promise for improving social justice and human and environmental wellbeing and there is currently a strong scientific, commercial, and political push towards improving the performance and expanding the deployment of predictive models in processes allocating benefits and harms. In this talk I argue that in order to deliver on its promises an interdisciplinary effort is needed to examine and assess the value-based judgments, which must be made at several junctions in the upstream design of ML models, which will have significant and differential impact on different groups and stakeholders.

Prof. Sune Holm earned his doctorate in philosophy from the University of St. Andrews, UK. He has worked and published on themes in the ethics of risk, bioethics, and philosophy of science. His current research concerns the ethics of AI with a particular focus on issues relating to algorithmic fairness and discrimination.