Bias in Data-Driven Artificial Intelligence Systems — An Introductory Survey

Artificial Intelligence (AI)-based systems are widely employed nowadays to makedecisions that have far-reaching impact on individuals and society. Their deci-sions might affect everyone, everywhere, and anytime, entailing concerns aboutpotential human rights issues. Therefore, it is necessary to move beyond tradi-tional AI algorithms optimized for predictive performance and embed ethical andlegal principles in their design, training, and deployment to ensure social goodwhile still benefiting from the huge potential of the AI technology. The goal ofthis survey is to provide a broad multidisciplinary overview of the area of bias inAI systems, focusing on technical challenges and solutions as well as to suggestnew research directions towards approaches well-grounded in a legal frame. In this survey, we focus on data-driven AI, as a large part of AI is powered nowa-days by (big) data and powerful machine learning algorithms. If otherwise notspecified, we use the general term bias to describe problems related to the gather-ing or processing of data that might result in prejudiced decisions on the bases of demographic features such as race, sex, and so forth.