Nordic states (except for Iceland) are often seen as trendsetters in AI with a very high level of AI readiness (Silo AI, 2021). One aspect of this has been their early adoption of national artificial intelligence strategies (NAISs). Sweden was the first to do so in 2018, followed by Denmark and Finland in 2019 and finally Norway in 2020. The NAISs provide a window through which to view a consolidated point where the Nordic states set out and emphasise ‘public sector decision-making in all aspects of artificial intelligence (AI) governance, from the regulation of third parties, to investments in national ecosystems and the direct adoption of AI in the public sectorʼ (Wilson, 2022, p. 2). As political imaginaries, NAISs frame the legitimacy of current political decisions and future developments for AI in society.

The article looks to the NAISs as visions for the future of healthcare in the Nordic region. Existing research has identified the importance of NAISs in the Nordic region (Tucker, 2023), and that more focus needs to be given to the cultural frames and values utilized (Robinson, 2020). As will be shown, the Nordic NAISs present a vision in which the expansion in the scale and scope of private-sector-driven AI applications in healthcare provision is inevitable, positive, and justified based on a promise of efficiency. More broadly, there is also concern that the NAISs are being used to establish neoliberal models as the only way to regulate AI for public goods, watering down alternative and non-market understandings of public participation and engagement in AI (Wilson, 2022).

Healthcare is particularly important, being seen as ‘an integral part of the welfare state model in all Nordic countriesʼ (Magnussen et al., 2009, p. 6). As the article outlines, healthcare has long since been shown to be tied closely to its societal context, both with respect to the relationship between political regime types and – more importantly – the impact of societal factors on human health. Healthcare, as discussed in the article, is a political field in which values and state–market relations impact which, and how, individuals are treated.

Applying awareness of healthʼs societal context to the Nordic NAISs, the article identifies the underlying values and social relations contained within. Importantly, as will be discussed, in the Nordic NAISs there is a predominance of market-oriented values in which private actors are seen to dominate, with healthcare turned into a consumer service. While the merits of alternate financing models can be debated, by adopting such a one-sided vision within the Nordic NAISs, the article argues that important political space is being foreclosed to the extent that the collective vision they contain can be fairly described as a politico-economic paradigm shift that is both publicly visible but made practically invisible through being framed with reference to abstract futuristic technologies. Noting this is important for understanding both the development of AI in healthcare within the Nordic countries as well as questioning why certain visions are excluded, and others made dominant.

The term ‘Nordic welfare modelʼ reflects that, when compared to other advanced economies, the Nordic states have similar social and healthcare policies. The model, which is rooted in a social democratic welfare state, ‘relies on the principle of universalism in promoting citizensʼ well-being and public policies emphasising social cohesionʼ (Kokko & Kork, 2021, p. 5). The narrative of the Nordic welfare model, which has been referred to as its ‘ethosʼ, speaks to an era where the welfare states were being guided by the principles of universalism, social resilience, and inclusion (Hellman et al., 2021, p. 125). This approach has positioned the Nordic states as prominent states utilising the development and implementation of public health policy focused on reducing societal inequalities within a tax-funded model (Raphael, 2014, p. 7). This can be understood by looking to the broader branding of the Nordic states internationally, what Browning (2007) refers to as ‘Nordicityʼ.

Yet, the public health sector in the Nordic states has not been immune to challenges. Since the 1990s many aspects of the Nordic welfare model have been replaced with market-oriented practices along the lines of ‘New Public Managementʼ (NPM). The adoption of NPM sits at odds with the domestic and international image of the Nordic states, where, as Kongshøj writes, what ‘has been distinctly Nordic […] is the degree to which local government (municipalities or counties) has been responsible for health care with some degree of state fundingʼ (2014, p. 25). However, both left- and right-wing governments have adopted such policies. The degree of implementation varies greatly, with, for example, Denmark acting only superficially, but Sweden incorporating NPM throughout its public decision-making systems (Hall, 2013). Overall, NPM reforms have shifted many aspects of Nordic public decision-making over to technocratic bodies and away from political accountability (Hall, 2013; Strange & Askanius, 2023). Stenius and Storbjork see the increase in market-oriented welfare regimes in the Nordic states as altering public procurement regulation to become a ‘struggle between the market logic and the welfare logicʼ (2020, p. 6).

One can see this very clearly in how the Nordic public health data, which was closely linked to the Nordic welfare state, is increasingly being imagined as a goldmine to be exploited for economic growth and the development of the private sector (Tupasela et al., 2020). Where there exists research on the datafication of public administrations in the Nordic region, we see little citizen engagement, but rather top-down technocratic and paternalistic processes (Broomfield & Reutter, 2022).

In the public arena, NPM reforms in healthcare have been framed as cost-saving and efficient, justified as necessary by an ageing population, the growing cost of healthcare, increasing expectations, and the individualisation of healthcare. This has led to national and regional variation and divergence in public health policy, including privatisation of aspects of healthcare. Private for-profit services have become increasingly prominent within the Nordic public healthcare setting, such as private residential care for children, but still the broader system is seen as ‘publicʼ and not for profit (Shanks et al., 2021, p. 128).

For political debates both within and about the Nordic countries, nevertheless, there is an overwhelming narrative that their welfare tradition remains strong. Almost all Nordic politicians frame themselves as publicly supportive of the Nordic welfare model, regardless of their policy positions, with many of the right-wing parties presenting economic reforms and anti-migrant policies as necessary to protect that welfare model. The NAISs provide a consolidated point of entry to see how these often contradictory narratives play out within the context of a future imaginary around AI.

Human healthcare is significantly determined by political factors (Braveman & Gottlieb, 2014). The relationship between living conditions, education level, and job precarity is well established (Benach et al., 2014). The withdrawal of such societal goods has also been used to undermine the life conditions intentionally for individuals due to, for example, their status as migrants as part of efforts to deter them from entering that society (Ruiz-Casares et al., 2013). The impact of residency status as well as marginalisation in healthcare was seen during the COVID-19 pandemic, where policies intended to limit societal access for key groups proved highly detrimental for efforts to control the virus (Dalingwater et al., 2023). Societal determinants of health are also often seen to cross generations with, for example, a womanʼs education level proving a key factor in the future health of her children, regardless of whether living in an advanced economy or developing state (Balaj et al., 2021).

Not only is human health impacted by societal determinants; the example of marginalisation shows how healthcare serves as an interlocutor between individuals and society. Healthcare is tied closely to societal types, with variations in financing – such as between private insurance-based compared to tax-funded models – often seen as synonymous with the stateʼs broader politico-economic character. Efforts to improve population health often overlap with broader social development policies. Health policies venture into themes such as ‘equalityʼ and ‘equityʼ as practitioners attempt to overcome the issues of inequality and injustice that limit healthcare access for marginalised groups (Rosas et al., 2022). Campaigns for universal health coverage have frequently encountered resistance due to the overlap with politico-economic issues, as seen markedly within the United States and long-stalled efforts to widen tax-funded healthcare (Birn & Nervi, 2019).

In several countries there has been criticism that individuals in need of healthcare are treated as ‘consumersʼ within a market model of healthcare, co-opting the patient to become a driver of a neoliberalisation of healthcare (El Enany et al., 2013). That shift has been exacerbated in several cases where AI has been implemented in healthcare, including issues such as data privacy, accountability, but most of all in core issues of equality and equity where there have been proven cases of developing bias and other problems potentially limiting access to healthcare (Norori et al., 2021). The visions through which AI is developed in healthcare shape not only the hospital context but speak to the notions of what constitutes a ‘goodʼ society, requiring critical scrutiny by which to identify the values framing those imaginaries and the power relations they embody.

Here the Nordic NAISs are approached in line with Wilson, identifying the principles, policy frames, and values they contain as ‘a consolidated point of reference for public sector decision-making in all aspects of AI governance, from the regulation of third parties, to investments in national ecosystems and the direct adoption of AI in the public sectorʼ (2022, p. 2).

The official English-language translations of the NAISs of Denmark, Finland, Norway, and Sweden were used.¹ These were officially translated by their respective states, enabling comparison through being in the same language. The documents included: Denmarkʼs NAIS, ‘National Strategy for Artificial Intelligenceʼ (Denmark, 2019); Finlandʼs NAIS, ‘Leading the Way into the Age of Artificial Intelligence: Final Report of Finlandʼs Artificial Intelligence Programme 2019ʼ (Finland, 2019); the Norwegian NAIS, ‘National Strategy for Artificial Intelligenceʼ, (Norway, 2020); and two documents for Sweden which collectively form the Swedish NAIS – the ‘National Approach to Artificial Intelligenceʼ, published by the Government Offices of Sweden in 2018 (Sweden, 2018), and the Swedish governmentʼs innovation agency Vinnovaʼs report titled ‘Artificial Intelligence in Swedish Business and Society – Analysis of Development and Potentialʼ (Vinnova, 2018).

To be clear, these are not legally binding policy documents, and, given the lack of existing national AI regulations, we cannot assess their policy impact. Further, they exist within socio-political and economic contexts, including that of digitalisation strategies at a national and regional level, and as such form just part of a wider nexus of social practices.

The article contends that they are important expressions of the statesʼ vision capturing the elite-level institutional discourse (Dexe & Franke, 2020). Through describing a future narrative of desire and aspiration the NAISs are what Jasanoff has called ‘socio-technical imaginariesʼ – ‘collectively held, institutionally stabilized, and publicly performed visions of desirable futures, animated by shared understandings of forms of social life and social order attainable through, and supportive of, advances in science and technologyʼ (2015, p. 3). Socio-technical imaginaries in the NAISs are important to explore as they not only shine a light on publicly performed visions of desirable futures, but they also normalise and legitimise certain current political choices (Ossewaarde & Gülenç, 2020) and fictional expectations in the economy (Beckert, 2013).

As established earlier, healthcare is highly political. Key to this is the issue of governance as well as the economic structures determining how individuals receive healthcare. Consequently, the analysis focuses on three core aspects when identifying how the NAISs narrate the role of AI in healthcare, these being: accountability and governance of AI in healthcare; the role of AI in healthcare; and the relationship between state and market within AI in healthcare.

The NAISs were coded using NVivo software focused on references to healthcare. This included the keywords: well-being, medicine, health(care), disease, illness, medical, diagnosis, cancer, nurse(s), physician(s), MR images, patient(s), and doctor(s). These codes were then put in a timeline, with only those referring to a future imaginary included in the analysis. To understand the underlying political content, the analysis focused on the three core aspects outlined above.

The NAISs considered here present four distinct but overlapping socio-technical imaginaries in which an AI future of healthcare necessitates significant regulatory changes. While Norway sees regulation as necessary on a proactive basis to help implement AI within healthcare, for the other states regulation is justified to reactively mitigate against risks to privacy, but, overall, is viewed as an obstacle to innovation and achieving the benefits of AI healthcare. Healthcare is not a politically neutral field but is structured along power relations. It is therefore significant that in all the NAISs the need to ensure transparency and uphold social values within an AI-oriented future healthcare rests on human physicians (Denmark), clinicians (Finland), healthcare workers (Norway) and a mix of private and public healthcare actors (Sweden). Norway stood out as the exception again where all other Nordic states represented prioritised the role of the private sector in guiding the development of a regulatory framework around AIʼs use in healthcare.

All four NAISs take a similarly broad approach to the future role of AI in healthcare, seeing it as including diagnoses, but with significant emphasis on its potential to help allocate resources more cost-effectively, as well as developing new ways to improve patientsʼ understanding of their own healthcare needs.

Given the Nordic welfare model and the strong role of the state this implies, it is significant that except for Norway all the Nordic states here see the future of AI healthcare as being largely dependent on the private sector, for its development and implementation at least, while its role as a service provider remains unclear. There is no consideration of the role the Nordic public sector plays in driving innovation for the public good, with all four NAISs articulating a vision in which only the private sector can produce the type of AI technology needed for healthcare. Norwayʼs exception concerns only whether regulation can help create the right conditions or is, as for the other states, an obstacle to progress. Importantly, all NAISs treat PPPs as normal and essential for the future of AI in healthcare. Like the findings of Tupasela et al. (2020), we can see that such partnerships are envisioned as predominantly focused on granting commercial access to public data, as well as allowing the private sector to use the public sector as a test bed for technological development. The relationships evident in these largely identical visions leave no space for imagining the public sector as an innovator or source of new technological development. Left only as a largely passive field, the value of the public sector rests on the extent to which it can be exploited by the private sector.

For the NAISs, the future of healthcare within the Nordic states is one in which the private sector is the provider. The public sector exists as only a regulator to resolve potential risks in much the same way as it does within retail transactions. The principal welfare benefits of AI in healthcare are presented as a potential to provide more detailed data on patients, theoretically enabling more precise healthcare guidance, monitoring, and allocation of resources. The strongly pronounced narrative that only the private sector can develop AI for healthcare contrasts with the absence of any description of the stateʼs role in funding healthcare, leaving the foundation of the welfare state uncertain. Also, while the vision of AI in healthcare in the four NAISs is framed positively as bringing better healthcare for individuals, it remains largely silent on how individuals can actively participate in their own well-being beyond their role as consumers. Overall, under the auspices of envisaging the future role of AI in healthcare, the Nordic NAISs analysed here present a strongly market-oriented policy vision in which welfare based on public provision has disappeared. While such marketisation is already evident in all Nordic states, the extent of that development as imagined in the NAISs is both radical and far removed from the common popular conception of the Nordic region as one based on welfare. The NAISs provide a means to articulate a vision that is both publicly visible but easily avoids wider scrutiny due to its technological futurism. The shift to the private sector envisaged in the NAISs is a significant paradigm shift with respect to altering the state-based model of Nordic welfare.

The four NAISs analysed in this paper as socio-technical imaginaries provide insight not into the future but, rather, present-day institutional values dominant within the Nordic region. Its key findings show that, despite a popular discourse in which the Nordic welfare model is supported against external threats, when imagining the future relationship between AI and healthcare – a cornerstone of welfare – the Nordic welfare model of state provision is largely absent. Furthermore, at their most extreme, the Nordic NAISs treat the state as little more than a reactive regulator, with healthcare viewed as an industry in which its future infrastructure – AI – will be provided by the private sector. This goes much further than the market-oriented model of NPM already established in the Nordic region, since it not only treats the public as ‘consumersʼ but also radically accelerates the shift towards for-profit actors as providers. Having established AI as not only a tool, but as the very infrastructure of healthcare in the future, the technology is then framed as something only the private sector can provide.

The vision presented in the NAISs is a paradigm shift compared to common self-perceptions of the Nordic welfare model, at least as expressed within political debates, and yet is both visible within the documents and avoids being the subject of public discussion due to the technological futurism of the NAISs. There is a clear distinction between the discourse within the NAISs and common perceptions of the Nordic states. To understand the extent of that shift requires tracing the longer-term implementation of those visions. In the meantime, inquiring into the values contained within documents such as the NAISs provides an often ignored means by which to trace changing notions of the ‘good societyʼ and, in this paper, has provided a valuable means by which to trace the changing politico-economic landscape of the Nordic region, as well as to study institutional discourse on the role of AI and healthcare.

The research was supported by co-financing between Malmö University and the Wallenberg AI, Autonomous Systems and Software Programme – Humanity and Society (WASP-HS).

On behalf of all authors, the corresponding author states that there is no conflict of interest. The authors have contributed equally to the article.