Are recent technological developments in artificial intelligence revolutionary, calamitous, something in-between? Are they inevitable, spontaneous, unpredictable? Jason Tucker examines how states actively shape developments in AI health. Their challenge is bringing the public back into decision-making in these developments
The narratives around the possibilities and dangers of AI are not new. We can categorise them broadly as falling into three camps: optimist, pessimist, and in-between. When it comes to AI in health, the optimists have centre stage.
They portray AI in health as overwhelmingly good for society and individuals. It will fix everything; it will completely revolutionise healthcare. We will have more time with clinicians, be better able to diagnose more diseases, do so earlier, faster, with more accuracy, and have precision treatment tailored to our individual needs. On top of that, healthcare will also be more cost effective and efficient.
This is a wonderful dream, but the actual clunky development of AI in health tells a very different story. Contrary to popular belief, AI development is not a free-for-all that will fix every problem given enough time and data. Rather, it is designed for very specific purposes and requires considerable resources to train it to become viable. Key actors prioritise certain AI applications and divert resources to their development, discarding other applications.
This narrows down the potential future of AI in health, based on various interests. It is, therefore, misleading to say that AI will revolutionise health. It can do so, in specific areas, but these will almost certainly be those in which key actors have chosen to focus their attention and resources.
AI development is resource intensive, and happens in specific, narrow ways as dictated by the priorities of certain key actors. So what are these priorities?
I’m interested in where states use public resources to facilitate the development of AI health. Which areas do they prioritise (or not)? And how do they sell these decisions to the public? Many states carefully nurture innovation ecosystems, fund public-private partnerships and cut or avoid implementing red tape. Certain futures of AI health are pushed forward, and others held back.
What concerns me is the limited space for the public to meaningfully participate and express their opinions on the future of AI health. My research and concerns centre on the future visions of AI health in the Nordic states.
The Nordic states are seen as leaders in AI, particularly in the health and MedTech sectors. They are also famous for their public healthcare, a cornerstone of Nordic welfare states. However, diminishing funding, an ageing population, and staff shortages are putting public healthcare under increasing strain. Some states have proposed AI to help them cope with these growing challenges.
To find out which AI health systems these countries are facilitating, and which they are not, we can turn to their National AI Strategies. These documents are fascinating resources; a consolidated point setting out public sector decision-making on AI in society. They offer an indication of funding and governance strategies, public-private partnerships, and priority areas. Within them we can identify socio-technical imaginaries of AI (including in relation to health) – future visions that justify policy past, present and future. Socio-technical imaginaries sit somewhere between policy and discourse, between hope and facts on the ground.
The futures outlined in Nordic states' National AI Strategies are in tension with those countries' public healthcare models, relying near-completely on the private sector
What the Nordic National AI Strategies tell us is surprising, and is in tension with those countries' public healthcare models. Across the Nordic states, there is a common future vision of the inevitable and increasing scale and scope of AI in health. The strategies frame these developments as almost completely reliant on private-sector innovation and implementation. Not only that, but they then use the promise of increasingly efficient healthcare provision to justify increasing private-sector dependency.
The Strategies contain only a few concrete examples of AI technologies. They are largely vague about what AI is, and how it will be implemented. This raises questions about transparency, accountability, and the role of values purportedly central to the Nordic public health model, like equality.
The narrowing of futures of AI health also raises concerns over public participation in decision making. The Strategies present these future visions of AI in health as inevitable, which leaves little space for debate.
But the Nordic states have backed themselves into a corner. While they support the development of certain AI technologies, they have little control over their process, outcome or implementation. Framing the public sector as unable to innovate renders these states dependent on the private sector to provide public health. Their only cards left are (de)regulation and granting private actors access to vast public health databases. Once these cards are played, states' role in healthcare governance, and space for meaningful public participation, is further diminished.
These AI strategies leave very little room for public participation, and even disempower the governments creating them
However, Nordic states have also promised certain future visons of AI health to the public that they must uphold. They must play these cards to make good on their promises. The states have framed themselves as simultaneously driving and being driven, caught in their own technological promise. The public are merely being taken along for the ride.
The public must be brought back into decision-making about the future of healthcare. This must be a future in which it is possible to reject or demand a range of AI technologies, and decide on the private sector's role in this. We can begin by foregrounding and critically interrogating Nordic states’ long-term strategies on AI health. By so doing, we can reintroduce meaningful participation, transparency and accountability into decision making on the future of AI health in the region.
Examining the pitfalls and challenges of AI integration in healthcare, this article serves as a timely reminder that technological advancements should be accompanied by robust ethical frameworks and inclusive strategies. While AI holds immense potential to revolutionize healthcare, it is crucial to address issues of privatization, accountability, and equitable access to ensure that the benefits are truly realized for all individuals and communities.