November 22nd, 2022 – 11:00 am
Algorithm Decision Systems are currently being designed to help humans in decision-making. For example, they are used to identify suspicious behaviours, detect hate speech or predict crimes. The impact of these systems on people and the society can be significant.
This seminar will talk about some of the privacy risks of these systems and will describe some recent attacks. It will also discuss why they sometimes fail to deliver. Finally, we will also show that, while evaluating these systems is essential, it is a very challenging task.
Bio:
Claude Castelluccia is a research director at Inria Grenoble and a founding-member of the Privatics Group. He is, since august 2021, Privacy commissioner at the CNIL (French Data Protection Agency) in charge of new technologies, digital administration, research and statistics. His research interests include data privacy, machine learning privacy, data anonymization and more generally privacy-preserving systems.
