Effectiveness dominates accountability in citizens’ perceptions of algorithms

Algorithms are increasingly used in the public sector in ways that affect people’s daily lives. While scholarly debate has generally emphasized the importance of accountable algorithms, it is in public sector organizations where accountability gains special importance. The present paper sheds light on how much value citizens place on relevant features of algorithms and especially on their accountability in relation to their effectiveness. It employs choice-based conjoint designs to present variants of algorithms used in policing and health care to respondents from a representative German sample. Two studies with overall more than 3000 participants consistently show that people are ready to trade away accountability features, such as stakeholder involvement and forms of transparency, for small gains in algorithm effectiveness. The findings have direct consequences for regulation, as citizens seem unlikely to demand accountable algorithms even in sensitive areas such as policing and health care