[Back]

S. Köszegi:
"Automated Decision Systems: Why Human Autonomy is at Stake";
in: "Collective Decisions: Theory, Algorithms And Decision Support Systems", 142; issued by: Springer Nature Switzerland AG; Springer Nature Switzerland AG, 6330 Cham, Switzerland, 2021, ISBN: 978-3-030-84996-2, 155 - 169.

For more than 50 years, humans have been using model- and data-based support systems in decision making with the hope that system supported decisions are not only better, but also more objective and fairer (i.e. more efficient and less biased). With data-based Artificial Intelligence (AI) systems, this hope is revived. AI systems are to support physicians in diagnosing illnesses, or assist managers when recruiting personnel. Furthermore, AI systems could even replace humans when deciding whether or not applicants receive a loan, an insurance policy or can lease a car. Even governments and public institutions are already making use of such automated decision systems including such diverse applications as setting bail in legal proceedings, risk assessment in youth welfare services or eligibility for state assistance to the unemployed and access to health services. The results of the current applications of automated decision systems are, however problematic. In this essay I examine the question of how automated decision systems impact the autonomy of humans and what requirements are to be placed on automated decision systems in order to protect individuals and society.

Human Autonomy; Automated Decision Systems