Public Administration and Computational Transparency in Algorithms - PACTA

Machine learning and associated technologies have progressed significantly over the last decade profoundly transforming our economy, our social relations, and our information retrieval practices. Now, these new technologies, with both their promise and their dangers, are being transferred into public administration in the form of algorithmic decision making (ADM). ADM is challenging traditional administrative practices by introducing an untested way of producing administrative decisions that have the ability affect the rights of individual citizens and questions the legality of the practice itself. PACTA investigates the fundamental problems relating to the interaction of legality in public administration and ADM. We do so by focusing on the following three areas of public administration::

a) welfare administration

b) policing/criminal investigation.

c) general problems in administrative law

To date, there is little to no research that explores where and how ADM is currently employed or how it is projected to be employed in the near future in Danish public administration. Most significantly, there is no research into the impact of ADM on legality or public trust. The PACTA research team is cross-disciplinary, involving pioneering scholars from law and computer science. In close collaboration, the team set out to fill this gap.

PACTA's goals are:

• to establish the breadth and depth of Danish public administrations’ use (and projected use) of ADM;
• to research the legality of ADM both theoretically as well as in actual implementation across various fields of administrationto research the role of trust in the implementation of ADM in public administration.

Future research is likely to build on the results from PACTA. This could be either in the production of legal guidebooks for governments or private businesses seeking advice in developing ADM software for decision making in a legal context or the creation of new models for including the public in the design of technological solutions adopted by administrative bodies.