2019-12-19 15:30 - 17:00 E1.14

Frederik’s talk concerns legal protection against discriminatory effects of algorithmic decision-making and other types of artificial intelligence (AI). AI can be used to predict who will commit crime, who will be a good employee, or who will default on a loan. The output of an AI system can seem rational and unbiased. But AI decision-making can also threaten the right to non-discrimination and other human rights. The talk gives examples of AI decisions with discriminatory effects. The talk also assesses which legal protection exists in Europe against discriminatory AI-driven decisions, and how that protection can be improved. Non-discrimination law, in particular through the legal concept of indirect discrimination, prohibits many types of AI-driven discrimination. Data protection law could also help to defend people against discrimination. However, the talk shows that both legal instruments have severe weaknesses when applied to AI. The talk gives suggestions on how legal protection against AI-driven discrimination could be improved.