Government by Algorithm - Protecting Democratic Values in the Deployment of Artificial Intelligence by the Public Sector.

The Norwegian public sector enjoys high levels of trust from citizens and is built upon a system where democratic values of freedom, equality and transparency are paramount. The central goal of the Norwegian Governments Digital Agenda is to realise ?a user-centric and effective public service?. The OECD in its Digital Government Review of Norway recommends that a data-driven approach be taken to realise this goal. They argue that this is ?an untapped opportunity? and that a ?sense of urgency? is lacking, which can prove detrimental to digital transformation of the sector. The realisation of this ?untapped opportunity? is now clearly on the agenda for Government, with a ?sense of urgency? becoming more apparent. The explosion in the adoption of Artificial Intelligence(AI) within the private sector exemplifies that whilst the benefits are potentially considerable, it is increasingly apparent that not all applications nor indeed the consequences are wholly desirable. Maximising the benefits and avoiding the pitfalls here has brought the challenge of ?algorithmic governance? to the fore. Whilst the public sector shares many governance challenges with the private sector, ethical and conscientious implementation of AI is more complex - given that many governmental decisions are especially weighty and democratically elected governments have special duties of accountability. Ensuring that democratic values are maintained, whilst reaping the benefits of ?data driven government? requires deep deliberation. The main objective of this project is to critically interrogate the potential benefits and pitfalls of Artificial Intelligence deployment in the Norwegian public sector and consider what appropriate governance measures could be employed. This will contribute to preserving Norways' democratic values and the high levels of trust it enjoys from citizens, in the scramble to realise a data-driven public sector.

-

The main objective of this project is to critically interrogate the potential benefits and pitfalls of Artificial Intelligence deployment in the Norwegian public sector and consider what appropriate governance measures could be employed. This will contribute to preserving Norways' democratic values and the high levels of trust it enjoys from citizens, in the scramble to realise a data-driven public sector. This will be realized by • establishing an overview and baseline for AI adoption in the Norwegian public sector. • testing whether the risks and challenges already identified in current research are also relevant to the Norwegian context and identify if there are others which should be considered. • going beyond a general public sector assessment by taking three case studies, to investigate if there is a difference in the perceived risks and challenges depending on the level of access to data and competence within the public sector or whether the private sector is involved. • identifying the most pressing issues and gaps in governance of AI deployment in the Norwegain context. This should also be transferable to the Nordic and broader international context. • identifying, critically evaluating and recommending potential remedies, including both soft and hard law tools. A multi-disciplinary approach will be taken, concentrating primarily on the legal, ethical and social challenges emerging in the adoption of this technology within the public sector. Research will be performed empirically and is inspired by critical data and algorithm studies. The project will begin by establishing the status within the public sector. It will take three case studies, namely a data & competence rich organization, a data & competence poor organization and a public-private partnership, to ascertain if risks and challenges differ depending upon the environment within which AI is deployed. It will then proceed to investigate and recommend governance remedies.