Automated Decision-Making Systems in the Public Sector – An Impact Assessment Tool for Public Authorities

How can we ensure a trustworthy use of automated decision-making systems (ADMS) in the public administration? AlgorithmWatch and AlgorithmWatch Switzerland developed a concrete and practicable impact assessment tool for ADMS in the public sector. This publication provides a framework ready to be implemented for the evaluation of specific ADMS by public authorities at different levels.

When deploying automated decision-making systems (ADMS) in the public sector, individual and societal trust in public authorities should be the ultimate benchmarks. Given the unique context in which public authorities act, the use of ADMS must be accompanied by a systematic assessment of potential ethical implications, ensuring transparency and accountability vis-à-vis those affected.

In the present report, we first outline the ethical foundation of our approach, from which specific questions for the evaluation of ADMS are derived. If the goal is to uphold the principles of harm prevention, autonomy, justice and fairness, and beneficence, there is no way around guaranteeing transparency, control, and accountability. The latter are not goals in themselves but unavoidable means toward guaranteeing an ethical use of ADMS in the public sector and for ensuring accountability. Based on that, we introduce a two-stage impact assessment procedure. It enables a triage of ADMS, indicating whether a system must be subject to additional transparency requirements. If this is the case, public authorities must provide a comprehensive transparency report. Lastly, we illustrate the use of the impact assessment tool by way of a fictional example.

The Impact Assessment Tool is a result of the study «Einsatz Künstlicher Intelligenz in der Verwaltung: rechtliche und ethische Fragen» (in German), on behalf of the Canton of Zurich. This was a collaboration of Prof. Dr. Nadja Braun Binder (University of Basel) and her team and AlgorithmWatch.

Ph.D. Michele Loi

Senior Research Advisor

Portrait Michele Loi
Michele Loi, Ph.D. is Senior Research Advisor at AlgorithmWatch Switzerland. He is Senior Researcher at the Institute of Biomedical Ethics and the History of Medicine, and Research Fellow at the Digital Society Initiative, University of Zürich. In 2015–2016, as a consultant, he coordinated the writing of the World Health Organization Guidance For Managing Ethical Issues In Infectious Disease Outbreaks and helped drafting the World Health Organization Guidelines on Ethical Issues in Public Health Surveillance.