Healthcare AI Policy in 2026: Only 7 of 38 OECD Countries Have an AI Strategy

[00:00:10] Dear listeners, welcome to Faces of Digital Health, a podcast about digital health and how healthcare systems around the world adopt technology with me, Tjasa Zajc. In 2026, emotions around health AI seemed mixed. What is clear is that the EU policy needs significant simplification and alignment built on trust.

[00:00:37] This was one of the conclusions at a recent conference on Responsible Scale of AI in Health organized by the Spanish Ministry of Health and OECD. Out of 38 OECD countries, only 7 have a formal AI strategy. Furthermore, only 11% have workforce upskilling programs and 9 conduct health technology assessment on AI,

[00:01:07] with 8 in the process of updating their HTA frameworks to evaluate AI's clinical relevance, cost-effectiveness, transparency and long-term safety. In today's episode, you are going to hear from a few speakers that were present at the conference in Madrid end of May 2026.

[00:01:31] Enjoy the show and if you haven't yet, make sure to subscribe to the podcast wherever you listen to your shows and also check out our newsletter, which you can find at FODH.substack.com. There's also a written version of the summary of the OECD conference there, along with a lot of content from the past. Now let's dive in today's discussion.

[00:02:09] Let's start with Eric Sutherland, Senior Economist at OECD, with a reflection on day one of the conference. I think we've heard very clearly that there is still a lot of work to do to collaborate together, to find the signal in the noise, to actually make sure that we are using our data for good

[00:02:32] and to measure the use of the data that we have because, frankly, we're not using it as efficiently and effectively as we could. And the third thing I would say is really that we have to focus on the regulatory environment. We have to simplify it, but we also have to make sure it delivers the needs of the public and of providers on a go-forward basis.

[00:02:58] Because ultimately, if we don't make sure that the public and providers trust what is being done with AI, then all the technology in the world isn't going to matter. An aspect that's usually not discussed enough, but was mentioned at the conference, is the impact of AI on youth.

[00:03:19] You're going to hear from Aferdita Mbitiki, Executive Director and Founding Partner of the Digital Transformation for Health Lab, and Erza Selmani, Research Fellow at the Digital Transformation for Health Lab. DTH Lab recently published research on shaping AI governance for young people.

[00:03:41] The DTH Lab, which is a global consortium, we are a research organization at the intersection of research, governance, and policy. And with that said, all of our work that we do, the core of it is the health and well-being of young children and adolescents. And we involve young people in all parts of the work we do. When we started looking at the analysis of digital national strategies, maybe 10 years ago,

[00:04:10] we noticed that youth health and well-being was largely omitted from those national digital health strategies. When we are now looking at AI governance policies and frameworks, we notice the same issues, the same gaps in the policies, where youth concerns, health concerns, and well-being are not addressed.

[00:04:35] Young people are only referred to in different strategies and documents as only as the future workforce, and not really for what is happening currently to them in terms of how AI systems are shaping their lives. In many countries, especially lower middle-income countries, you have countries that have 50-60% of the population, under 35 years old.

[00:05:02] How do you build a digital health strategy or an AI framework or policies, disregarding the needs of 50-60% of your population? Valentina Stramelo, executive director of the European Patients Forum, presented some results from a study and survey that EPF did in 2025. Based on the results from 1,000 patients that responded to the survey,

[00:05:30] here is where the sentiment lies when it comes to patients and AI. The main findings include patients tend to be very positive about the potential of AI in healthcare, to the point that nearly 98% of them responded positively about AI.

[00:05:51] Of course, they also have some concerns, mostly because they are afraid of potential biases that AI can lead to. One of the takeaways was also the importance of patient information. Patients would like to receive direct communication from healthcare professionals, and also they would like to receive, if possible, AI-specific consent forms.

[00:06:19] Patient information is key to support the trustworthiness of AI from a patient perspective. So that's an element that should be taken into account without scaring patients. So we need to strike the right balance in terms of which level of communication should be shared with patients and how detailed it should be. But it's important to bring transparency.

[00:06:44] Last but not least, also a takeaway was that still patients see healthcare professionals as the key point of reference when it comes to the management of their health condition. And so they would still refer to healthcare professionals as the key point of contact and of information where AI can be a tool to complement the information they can receive from healthcare professionals.

[00:07:13] So how should policy move next? Dr. Ricardo Baptista Leite, CEO of HealthAI, the Global Agency for Responsible AI in Health, outlined five recommendations for EU policy. HealthAI has just published an exciting report with the World Economic Forum focusing on how we can purpose the EU AI Act for health and economic competitiveness.

[00:07:40] And we have five key recommendations on how we can do so more effectively moving forward. Recommendation number one is that countries need to address the issue that they have dual supervisory approaches that need to be coordinated. You have horizontal AI supervision, but also vertical AI supervision in the specific field of health.

[00:08:04] Countries need to perhaps look at the opportunity to do joint mock exercises to see how that coordination can be done, to make sure that the different parts of the governmental institutions can work in a coordinated fashion to ensure that the technology is being used to improve health for all. Recommendation number two, the European Commission but also the member states need to solve the issue of the barriers in terms of access to markets,

[00:08:33] also for small and medium enterprises that are developing innovation. And that means fixing several issues around the governance pathway, but specifically health technology assessment and reimbursement. If we do not fix this issue, scalability will not happen. The technology will not be accessible to everyone who needs it. Recommendation number three is how the European Commission needs to ensure that there's an interplay that is functioning

[00:08:59] between different pieces of legislation and frameworks that already exist in the European Union. That includes the EUA-AI Act, MDR legislation, the European health data space, and making sure that the different directorate generals within the European Commission are working together and learning from what countries are doing so that effectively these different frameworks do not create more complexity, but actually accelerate the responsible adoption of technology and the scalability of technology within the European Union.

[00:09:30] Recommendation number four, the European Commission should be able to play a role coordinating lessons learned, good and bad, from what is happening at member state level. If we're able to maximize what is happening in each country, we will be able to learn faster, to improve, and also to have lessons learned so that we don't repeat mistakes.

[00:09:52] We believe that the DigiSanté and the AI office within the European Commission should coordinate this effort with the collaboration, of course, of the member states. Finally, recommendation number five, probably the most important of them all, it's around stakeholder engagement. We need to make sure that all stakeholders are involved in the process from legislation to the development and deployment and implementation of technology.

[00:10:22] And that means not just the technology companies, not just governments, but also those who will be the potential beneficiaries of AI for health. And that includes, of course, of course, and patient organizations who need to be included not as tokens but as active actors in the decision-making process. If we are serious about scaling technology in a way that is truly equitable, that means we need to include everyone to ensure that AI can improve health for everyone everywhere.

[00:10:53] The final thought comes from Persephone Dupi, senior medical officer at the Finnish Institute for Health and Welfare and the president of the European Federation for Medical Informatics. As she emphasizes, digital literacy is something that we need to address more.

[00:11:13] We have had very interesting discussions. What has been confirmed is that what we are doing in the EU level, trying to leverage the HDS framework and all the remaining components of the EU data strategy, like the MDR, IVDR revision, the AI Act and all the rest of legal efforts to improve the data infrastructure and quality of data and data sharing is very much into the center of these discussions we are having.

[00:11:44] As a medical informatician and representing the European Federation for Medical Informatics, I would say that we need to put a lot of emphasis in supporting AI literacy, both for the healthcare professionals and the public. And we certainly have activities for that. And as a public health specialist, I would say that was the component that was least addressed.

[00:12:04] And I hope there will be more emphasis in the future, because we really need to leverage these tools for prevention and for shifting the weight in healthcare services from treatments to actually supporting healthy lifestyles and citizens' well-being.

[00:12:45] Thank you.