Behavioral Consequences of AI, Big Data and Algorithmic Decision-Making in Public Services

Abstract

The digitalization of public services is in full swing. Specifically, machine-learning technologies and the algorithms that power them hold a huge potential to make government services fairer and more effective, ‘freeing’ decision-making from human subjectivity enhance the efficiency of public services, reduce administrative burdens, and lower personnel costs. Algorithms today are used everywhere from welfare to criminal justice; for instance, they can predict recidivism better than criminal court judges. Research indicates that the digitization of public services caused by the introduction of algorithms may cause profound shifts in the way bureaucrats make decisions. Overall, using such technological systems in administrative processes and public service delivery is supposed to be beneficial, yet little is known about the effects of those changes on micro-level attitudes and behavior of public servants and citizens. This panel invites papers advancing theory, providing empirical insights, or speaking to public management practice in terms of behavioral consequences of digitalized public services. Any papers applying experimental (lab, field, survey) methods, other quantitative, or qualitative research approaches are welcome. Examples of questions that we would like to explore in this panel are: • Does algorithmic decision-making lead to more or less bias in individual decision-making? • Are algorithmic/digitalized services perceived as more or less trustworthy by citizens? • In what ways do technologies such as big data, AI and algorithms alter the way bureaucrats make decisions? • Do algorithmic/digitalized citizen-state encounters affect how citizens perceive public service performance? • What are the antecedents or conditions driving citizens’ satisfaction with algorithmic/digitalized public services? • Other questions taking a micro-level perspective on the effects of digitalization of public services

Title	Details
The Causal Effect of Digital Interaction and Service Failure on Citizens’ Satisfaction with Routinized Public Services. Evidence from a Vignette Experiment	View Paper Details
The Impact of Sector Difference on People’s Data Sharing Intention: an Experimental Study	View Paper Details
Automation Bias in Public Policy? Assessing Decision-Makers’ Overreliance on Algorithmic Advice Via a Survey Experiment	View Paper Details
Blame Avoidance Strategies and the Use of Algorithms in Public Policy Decision Making. A Case Study on Criminal Justice Policies in the US	View Paper Details
A Machine Learning Approach to Public Comments for Regulatory Policymaking	View Paper Details