Algorithms increasingly influence organizations in the public sector. Even though algorithms can strengthen the effectiveness and efficiency of the public sector, there is also reason for concern. Can we trust that these algorithms respect our privacy and work in a non-discriminatory manner? Researchers and citizen organizations call for algorithmic transparency since they expect that this will result in justified trust in the use of these algorithms. This research project investigates this claim through ethical research (mapping ethical dilemmas in algorithmic transparency for justified trust) and experimental research (understanding empirical relations between algorithmic transparency and trust) and design research (developing design principles for algorithmic transparency). The research objectives are to develop a thorough ethical and empirical understanding of algorithmic transparency and use this understanding to develop design principles to guide the development and implementation of (1) predictive policing algorithms for operational planning and (2) intake systems for crime reporting by citizens. We will focus specifically on the police since they are quite advanced in the use of these algorithms and this usage generates specific concerns in view of the key position of the police in dealing with conflict in our society. Our research combines knowledge from the humanities about ethics and ethical dilemma’s with knowledge from social science on empirical relations between transparency and trust. For the development of the design principles, we will also build upon knowledge from information systems research to ensure the validity and usefulness of these principles for designing algorithms.
PERITIA – Policy, Expertise and Trust in Action – is een Europees project (Horizon 2020) dat onderzoek doet naar de condities waaronder mensen experts vertrouwen bij het vormgeven van publiek beleid. Het UU-deel van dit onderzoek richt zich op vertrouwen in expertise in een sterk veranderend media en informatie landschap.