A Systematic Assessment of National Artificial Intelligence Policies: Perspectives from the Nordics and Beyond

Abstract

Echoing the evolving interest and impact of artificial intelligence on society, governments are increasingly looking for ways to strategically position themselves as both innovators and regulators in this new domain. One of the most explicit and accessible ways in which governments outline these plans is through national strategy and policy documents. We follow a systematic search strategy to identify national AI policy documents across twenty-five countries. Through an analysis of these documents, including topic modelling, clustering, and reverse topic-search, we provide an overview of the topics discussed in national AI policies and contrast the differences between countries. Furthermore, we analyse the frequency of eleven ethical principles across our corpus. Our paper outlines implications of the differences between geographical and cultural clusters in relation to the future development of artificial intelligence applications. CCS CONCEPTS • Human-centered computing → Human computer interaction (HCI); HCI theory, concepts and models; • Social and professional topics → Government technology policy.

Designing Visual Markers for Continuous Artificial Intelligence Support

“…The growing use of AI across a range of application domains has led to an increased focus on usable [9,39], fair [49,53], and transparent [4] systems within the HCI community as well as the wider public debate [33,50]. The successful uptake and deployment of these algorithmic systems does, however, rely on the incorporation of stakeholders' knowledge and feedback [62].…”

Effect of Information Presentation on Fairness Perceptions of Machine Learning Predictors

“…Technological developments have led to an increase in the deployment of artifcial intelligence (AI) based decision-support solutions. Having moved beyond theoretical solutions or theories, such systems now have a profound efect on both individuals and society at large [25,57]. The use of AI spans across a variety of application domains, including high-impact areas such as loan approval [16,36], policing and law enforcement [29,31], and hiring processes [14,46].…”

“…Our study was limited to American crowdworkers, and prior work shows that perceptions towards e.g. AI behaviour and fairness difer signifcantly between geographical and cultural clusters[3,57].…”

Artificial intelligence in human factors and ergonomics: an overview of the current state of research

“…Buxmann and Schmidt [6] Labor market BMBF [57] Labor market Frey and Osborne [58] Labor market Grace et al [59] Labor market PWC [53] Labor market Economic production factor Begleitforschung PAiCE [54] Economic production Factor Nature [55] Economic production factor LBBW Research [56] Economic production factor Nankervis et al [37] Production factor Hartmann et al [60] Production factor Ansari [16] Vocational training concepts Ansari and Seidenberg [17] Vocational training concepts Feigh et al [26] Vocational training concepts Deutscher Bundestag [41] Legislation and politics Deutsche Bundesregierung [42] Legislation and politics Cath et al [43] Legislation and politics Singh [45] Legislation and politics van Berkel et al [44] Legislation and politics OECD [46] Legislation and politics Clarke [48] Legislation and politics Steels and López de Mantaras [51] Legislation and politics Park [47] Legislation O'Sullivan and Thierer [52] Politics van Nuenen et al [49] Discrimination Santow [50] Discrimination…”

“…The study by van Berkel et al compared the national policy documents on AI of 25 countries. It was shown that the following topics occupy the world[44]:…”