Democratic behavior

Public Fairness Perceptions of Algorithmic Governance

This proposal aims to study what fairness perceptions citizens adhere to in relation to governance decisions based on algorithmic information processing, and how its use may affect democratic legitimacy. Such procedures are increasingly being introduced by government institutions to help making decisions that impact individual citizens on topics such as giving defendants parole, approving immigration applications, and determining eligibility for welfare programs. Thus, we are on the verge of a revolution in public sector decision-making processes, where computers will take over many of the governance tasks previously assigned to human bureaucrats. With it, the conditions for impartial and transparent treatment of citizens are changing. Increased capacity to process relevant information enhances the potential for making more accurate and efficient judgments. Yet, we also run the risk of creating a black box society where citizens are being kept in the dark about the decision-making processes that affect their lives, potentially undermining the legitimacy of governmental institutions among the citizens they serve. While significant attention in the recent few years has been devoted to normative discussions on fairness, accountability, and transparency related to algorithmic decision making, little is still known about citizens’ views on this issue. There is thus an imminent need to study these emerging governance developments from a political science perspective. This proposal aims to fill this gap by organizing both in-depth group discussions on this topic among representative samples of the Norwegian population through deliberative polling, as well as conduct survey experiments on larger representative survey samples. The Center for Deliberative Democracy at Stanford University and The Digital Social Science Core Facilities (DIGSSCORE) at the University of Bergen are involved in the project, ensuring high data quality from top social science infrastructures.

Demokratiske algoritmer: Hvordan oppnå legitimitet og rettferdighet i automatiserte beslutningsprosesser i offentlig forvaltning

Den pågående automatiseringen av beslutningsprosesser i offentlig forvaltning representerer en omveltning innenfor byråkratisk myndighetsutøvelse. Tilgang på store mengder relevant digital data og økende muligheter for å behandle informasjonen gjør at oppgaver som tidligere måtte behandles manuelt kan overlates til hel- eller halvautomatiserte prosesser med vesentlig redusert menneskelig inngripen. På den ene siden gir denne utviklingen store effektiviseringsmuligheter og potensial for offentlige besparelser. På den andre siden er ivaretakelsen av forvaltningens legitimitet i befolkningen et risikoaspekt i denne utviklingen. Det overordnede målet med det foreslåtte prosjektet er å kartlegge ut fra et demokratiperspektiv om, og i så fall hvordan, oppfattelsen av NAV som institusjon blant innbyggere i Norge påvirkes av en overgang til økt grad av automatisert saksbehandling. For å besvare forskningsspørsmålene vil vi våren 2021 gjennomføre en spørreundersøkelse i Norsk medborgerpanel, som er et befolkningsrepresentativt panel som samler inn data til forskningsformål.

Procedural Legitimacy

The primary scientific objective of the PROLEG project is to better understand how democratic institutions and decision-making bodies should organize decision-making procedures and implementation procedures in order to make them more legitimate in the eyes of the public. We study if and how variations in political decision making procedures can make the outcomes more acceptable to the citizens, and especially to those who disagree with the outcome. Do people share universal perceptions of fair decision makingn procedures? In a nutshell, the PROLEG project will address this issue and generate new knowledge that can be used to improve governance in the future. This will be accomplished by conducting experimental and observational studies on the mechanisms of accepting decision-making procedures. The data will mainly be generated within the infrastructure of DIGSSCORE at the University of Bergen, Norway, taking advantage of changes in technology and research methodology that combine to bring computer laboratory research and survey studies closer together.