Introduction

Although the algorithms of machine-learning methods have brought issues of discrimination and fairness back to the forefront, these topics have been the subject of an extensive body of literature over the past decades. But dealing with discrimination in insurance is fundamentally an ill-defined, unsolvable problem. Nevertheless, we try to connect the dots, to explain different perspectives, going back to the legal, philosophical, and economic approaches to discrimination, before discussing the so-called concept of “actuarial fairness.” We offer some definitions, an overview of the book, as well as the datasets used in the illustrative examples throughout the chapters.