Cracking the Code of Digital Health / DAVOS 2025

Summary

This panel discussion focused on the current state and future potential of digital health and AI in healthcare systems globally. The panelists, representing various healthcare organizations and technology companies, explored the challenges and opportunities in implementing digital health solutions and AI across different countries and healthcare systems.

Key points included the need for better data liquidity and interoperability to fully leverage AI’s potential in healthcare. Panelists emphasized that while progress has been made, there is still a significant gap between available AI technologies and their widespread adoption in healthcare settings. They discussed barriers such as regulatory challenges, trust issues, and the need for healthcare systems to adapt their infrastructure to accommodate new technologies.

The importance of public-private partnerships and international collaboration was highlighted as crucial for advancing digital health initiatives. Panelists also stressed the need for responsible AI development, balancing innovation with patient safety and trust. The discussion touched on the potential for AI to address healthcare worker shortages and improve access to care in underserved areas.

Challenges in data standardization, privacy concerns, and the need for better integration of AI tools into existing healthcare workflows were identified as key areas for improvement. The panelists agreed that while AI has the potential to significantly improve healthcare outcomes and efficiency, its implementation must be done thoughtfully and with consideration for ethical implications and equitable access.

The discussion concluded with calls for increased investment in digital health infrastructure, better training and education for healthcare professionals in AI technologies, and the development of clear pathways for countries to adopt and scale digital health solutions. Overall, the panel emphasized the transformative potential of digital health and AI in addressing global healthcare challenges, while acknowledging the complexities involved in their widespread adoption.

Keypoints

Major discussion points:

– The current state of digital health and AI adoption in healthcare systems globally

– Challenges and enablers for scaling digital health solutions, including data sharing, trust, and regulatory issues

– The potential for AI and digital tools to improve health outcomes and address healthcare workforce shortages

– The need for responsible development and deployment of AI in healthcare

– Opportunities for public-private partnerships and international collaboration to advance digital health

Overall purpose/goal:

The discussion aimed to assess the progress of digital health and AI adoption in healthcare, identify key barriers and enablers, and explore potential solutions and collaborations to accelerate the responsible implementation of these technologies to improve global health outcomes.

Tone:

The overall tone was constructive and forward-looking, with panelists acknowledging challenges but expressing optimism about the potential of digital health and AI. There was a sense of urgency to accelerate adoption while ensuring responsible development. The tone became more action-oriented towards the end as panelists suggested concrete next steps.

Speakers

– Karen Tso: Anchor of Squawk Box Europe on CNBC

– Gianrico Farrugia: President and Chief Executive Officer, Mayo Clinic

– Roy Jakobs: President and Chief Executive Officer, Royal Philips

– Shobana Kamineni: Executive Chairperson, Apollo Health Company

– Nikolaj Gilbert: President and Chief Executive Officer, PATH

Additional speakers:

– Audience member: Asked a question about countries collecting health data centrally

Digital Health and AI in Healthcare: Challenges and Opportunities

This panel discussion, moderated by Karen Tso, featured leaders from prominent healthcare organizations and technology companies exploring the current state and future potential of digital health and artificial intelligence (AI) in global healthcare systems. The conversation highlighted significant challenges in implementing these technologies at scale while emphasizing their transformative potential for improving healthcare outcomes worldwide.

Current State of Digital Health and AI Adoption

The panelists agreed that despite the immense potential of AI in healthcare, there is a widening gap between available AI technologies and their widespread adoption in healthcare settings. Roy Jakobs of Royal Philips noted that digital health adoption is occurring in pockets but not yet at scale. This sentiment was echoed by Gianrico Farrugia of Mayo Clinic, who observed that the gap between AI capabilities and healthcare adoption has recently widened.

The discussion revealed that even in technologically advanced countries like the United States, digital health solutions are not widely adopted despite significant healthcare spending. Roy Jakobs provided a striking example, stating that only 10% of pathology is currently digitized, which severely limits AI applications in this field.

Barriers to Digital Health Adoption

Several key barriers to the widespread adoption of digital health and AI were identified:

1. Antiquated Healthcare Architecture: Gianrico Farrugia emphasized that current healthcare systems are not optimized for AI integration, describing the existing architecture as antiquated and ill-suited for modern technologies.

2. Data Liquidity and Interoperability: Roy Jakobs highlighted the lack of data liquidity and interoperability between systems as a significant obstacle. This view was supported by Nikolaj Gilbert of PATH, who stressed the need for improved standardization and interoperability of health IT systems.

3. Trust Issues: Both providers and patients often lack trust in AI systems, as noted by Gianrico Farrugia. This underscores the need for responsible AI development and clear communication about its capabilities and limitations.

4. Regulatory Challenges: Roy Jakobs pointed out that regulatory hurdles around data sharing and AI in healthcare continue to impede progress.

5. Training and Education: The need for better training of AI algorithms on clinical data was emphasized, along with the importance of upskilling and reskilling the healthcare workforce on digital technologies, as mentioned by Shobana Kamineni of Apollo Health Company.

Potential Benefits and Applications

Despite these challenges, the panelists were optimistic about the potential benefits of digital health and AI:

1. Improved Healthcare Outcomes: Gianrico Farrugia asserted that AI can improve diagnosis, treatment, and prevention of diseases in ways that were not possible even three years ago. He highlighted Mayo Clinic’s MiracleLink platform, which has a global reach and includes specific AI applications in cardiology, such as detecting heart failure with preserved ejection fraction.

2. Increased Access to Care: Shobana Kamineni highlighted how digital health can increase access to care, especially in rural areas, and enable low-cost preventive care at scale. She shared that in India, digital dispensaries are being used to provide healthcare access to remote areas, and preventive healthcare can be delivered for as little as $8 per year.

3. Enhanced Productivity: Roy Jakobs noted that AI can enhance the productivity of healthcare workers, potentially addressing workforce shortages, a point also made by Nikolaj Gilbert.

4. Global Health Advancement: Nikolaj Gilbert emphasized the potential for digital health to advance Sustainable Development Goals in developing countries. He mentioned an ongoing pilot project testing AI in clinical decision-making in three African countries.

Strategies for Advancing Digital Health Adoption

The panelists proposed several strategies to accelerate the adoption of digital health and AI:

1. Systems Approach: Roy Jakobs emphasized the need for a systems approach in healthcare, involving technology, clinical practice, financial systems, and regulatory aspects.

2. New Healthcare Architecture: Gianrico Farrugia called for the creation of a new healthcare architecture better suited for AI integration.

3. Public-Private Partnerships: The importance of developing public-private partnerships to validate AI algorithms was stressed.

4. Building Trust: Roy Jakobs emphasized the need to focus on building trust through responsible AI development, noting that AI should not be held to higher standards than human doctors.

5. Workforce Development: Shobana Kamineni highlighted the importance of upskilling and reskilling the healthcare workforce on digital technologies.

6. Standardization and Interoperability: Nikolaj Gilbert advocated for improved standardization and interoperability of health IT systems, particularly in the digitalization of public healthcare systems in the global south.

Global Perspectives

The discussion touched on varying global perspectives on digital health adoption:

1. India: Shobana Kamineni described how India is leveraging digital health to improve access for 1.5 billion people, including the use of digital dispensaries in rural areas.

2. United States: While leading in AI technology development, the US is lagging in healthcare applications, a point that sparked some debate among the panelists.

3. Singapore: Roy Jakobs mentioned that some countries, like Singapore, are advancing quickly in digital health adoption.

4. Developing Countries: Nikolaj Gilbert emphasized the need to consider country-specific contexts for AI deployment and the potential for digital health to advance healthcare in developing nations.

The panel also briefly addressed an audience question about countries collecting centralized health data, with panelists discussing the potential benefits and challenges of such approaches.

Conclusion

As the discussion concluded, panelists offered their thoughts on priorities for the next 12 months to advance digital health. These included focusing on specific use cases, improving data interoperability, and accelerating the adoption of proven technologies.

The panel discussion highlighted the complex landscape of digital health and AI adoption in healthcare. While there was consensus on the transformative potential of these technologies to improve healthcare outcomes, increase efficiency, and address workforce shortages, significant challenges remain. These include structural barriers in healthcare systems, data sharing and privacy concerns, and the need to build trust among both providers and patients.

The panelists called for increased investment in digital health infrastructure, better training and education for healthcare professionals in AI technologies, and the development of clear pathways for countries to adopt and scale digital health solutions. They emphasized the importance of public-private partnerships and international collaboration in advancing these goals.

As the discussion concluded, it was clear that while the road to widespread adoption of digital health and AI in healthcare is complex, it is also filled with immense potential to transform global health outcomes. The key lies in addressing the identified challenges through collaborative, innovative, and responsible approaches that consider the unique contexts of different healthcare systems worldwide.

Karen Tso: Distinguished guests, ladies and gentlemen, thank you so much for joining us. I’m Karin Cho, anchor of Squawk Box Europe on CNBC, and it’s such a pleasure to be here with you today. Now, let me just take you back over history. I remember being here a few years back and we were asking all the usual questions. Outlook, geopolitics, macro challenges, what lies ahead, a little helicopter view. We started interviewing leaders from Asian businesses and they started fleshing out to us this storytelling about how they’re dealing with this thing called COVID, that genuine cases on the ground, highly contagious. And, of course, we all know what transpired after that. We went into lockdown. Weeks and months later, across other countries, they, of course, started talking about COVID vaccines. And fast forward, we were talking about apps to track some of the virus and the pandemic, who had had vaccines, who hadn’t. Vaccine certificates went online. And, of course, we saw stretched healthcare systems also pivot, trying to use their resources through digital technology. It was a new dawn for the industry, being at the forefront of capitalism. I can tell you that we certainly saw private and public money pivot into this space. But fast forward to 2025, what happened to that momentum? Where is it? Healthcare systems, governments, we know, are trying to put a lid on fiscal spending, so you would think there would be appetite, motivation for digital as a solution. Does AI bring fresh opportunities in 2025? Well, I know the World Economic Forum Centre for Health and Healthcare has been hard at work on this issue with the Digital Healthcare Transformation Initiative, and there’s plenty on this that you can read online. But let me introduce you to your panellists, because we do have terrific speakers. Apparently, they’ve spent most of Davos together, so they’re going to have a lot of insights to share with us. I’m pleased to welcome to the stage Dr Gianrico Ferrico, who is President and Chief Executive Officer, Mayo Clinic. Also joining us, Roy Yacobs, President and Chief Executive Officer at Royal Philips. Shona Kamineni, who is the Executive Chairperson, Apollo Health Company. And Nikolaj Gilbert, who is President and Chief Executive Officer at PATH. Thank you all so much for joining us here for this Cracking the Code of Digital Health session. Shonrika, let me start with you. There is a stat. In the United States, the biggest healthcare system, spending reached nearly $5 trillion for the first time. $5 trillion. And your clinic plays a leading role with major diseases and the treatment of them, from cancer to cardioproblems, diabetes. And you’ve also stepped up digital transformation. Now, the US is not alone in facing this challenge of rising costs, escalating costs. What’s your assessment now as to why countries haven’t turned to technology, where they’re at in using digital and AI when it comes to treating health? Where are we at and where do we need to lift our game?

Gianrico Farrugia: Well, thank you for moderating this panel. And I do want to thank WEF for doing it too, because my view is that whether or not we’re stuck in what you just described, a healthcare system that is frustrating, and whether instead we can transition to a healthcare system as more accessible, scalable, and gives better outcomes, to me, is highly dependent on how we deal in becoming digitally first in healthcare. Now, the good news is that we’re further along than most of us thought we would be. I come from a system where we’ve pushed digital and AI into healthcare for a while. But even I’m surprised that at Mayo Clinic, we have 320 AI algorithms in the practice every day. Last year, we did over 800,000 digital visits. The issue is that while that’s happening in most healthcare organizations, three, six months ago, there was a further acceleration AI, and healthcare did not accelerate at the same rate. So indeed, the gap has widened. The gap shouldn’t be there, because if you look at it objectively, we know that healthcare plus AI is better than healthcare on its own. We know that you can diagnose, treat, and even prevent disease in ways you couldn’t do three years ago. So why the gap? It’s because at the moment, healthcare organizations, governments are unable to take what’s currently available and meet the needs of what populations expect from them. And I believe that gap is structural. That gap is because we’re trying to overlay AI and digital healthcare onto an architecture for healthcare that is antiquated and doesn’t work. And so what we need to do is take a step back and create a new architecture. To give a concrete example, six years ago, we did this. We established the MiracleLink platform. It now has over four continents. It has 61 healthcare provider organizations on it, 81 solution developers. And at Davos, three years ago, I made the point that we can use a lateral cardiogram to diagnose heart failure by predicting ejection fraction. We placed that on the platform, and since then, there are now 14 different cardiac diagnoses we make every day by AI on our patients. But because we made it available to others, nurses can now use it, and nurses plus the algorithms are better than cardiologists without it. Taking one of those algorithms and placing it into a stethoscope, nurses in Nigeria have been able to reduce mortality from pregnant women. And that algorithm taken by others and making it country appropriate is being used now to diagnose heart failure in kids in Peru. So the point is that with the right architecture, we can indeed scale AI at a very low cost. And what we need to do now is create those mechanisms to allow countries, organizations… organizations to come together and create both the roadmap as well as the financing that’s responsible, that is needed in order to make that switch. There are of course other issues, trust I’m sure you’ll get to, but I’ll end in a very positive note. For the first time I can remember in 35 years in healthcare, we are not only now talking about the problems in healthcare, we have defined solutions that are scalable. So shame on us, shame on all of us, shame on government, if we cannot at this moment in time come together and create the pathways and the architecture to be able to do what

Karen Tso: we already know we can do, provide better outcomes for patients at a scale that was unimaginable a few years ago. Gianrico, thank you, you’ve given us a lot to think about already. Roy, let me turn to you because one of the comments I hear most often when I attend technology events is that look, the area that AI will disrupt the most is going to be healthcare. So where is it?

Roy Jakobs: I mean, you’ve been working on this, I’ve got to say, but we haven’t seen the system-wide impact when it comes to some of those digital solutions, the scale that is sorely needed from here. So tell us your perspective, you see the current wave of technology, you’ve got all the pictures coming through from some of those unicorns as well, where can we see some of these solutions best deployed to have impact? Yeah, great question and I will probably take it from two sides how I look at it. One is the glass half full and the glass half empty. Now if you look at it from a negative side, you indeed say kind of the system is struggling to get the full benefits of AI because the adoption of AI is not easy. And that is already not easy within a certain practice of healthcare, because as you know there’s a lot of different practices in healthcare, let alone if you need to do that system scale. And that’s what we need to talk about, how can we actually get healthcare moving at a system scale? And I always say when you need to get the system moving, you need to get the different part moving at the same time. So yes, technology has arrived and is ready to be scaled. But we know as a technology company that technology alone can never do it. to do the job. It takes the clinical practice that actually appreciates what it can do for them. And actually also it’s made easy to work with it in order to really get the benefits and then scale it. But even more than that, you then also need to get the financial system to reward that new practice. Because if you still keep rewarding the old practice, why would you adopt a new technology? Actually, it’s more painful. You don’t get the incentive for it. And then the regulatory side is also very important. And especially if you talk AI, you talk data, and we know where we are with regulatory on data, in particular data liquidity. So I think when we look at how can we move at scale, we need to get the system moving at the same time. Now at COVID, and you referred to it, we were able to do that because we were single-minded focusing on solving a major crisis that the whole world was rallying around. And then you saw that technology was brought to bear, systems were willing and able to change in weeks or days. They were mobilizing massively. Regulators were kind of adopting regulation to accommodate that. And actually, financial schemes were changed in a day or a week to actually support that. So certainly everything was possible, but it needed that level of crisis to get the system changing. So whilst we’re talking about a healthcare crisis, maybe we don’t actually act on it as it is in crisis. And that’s, I think, what we see at a systems level. And it’s a complex system. So that’s when I say, okay, so there’s a lot to be done to get us moving at a systems level to get all these different actors lining up at the same way. Now on the positive side, I do see that there is adoption in a lot of the practices. And if I look at right and left, actually we have some leading examples, but also I see it on a country level. If I look to Singapore, I see it in the Nordics, in Europe, I see it in someone in the Asian countries where you see practice really reinventing themselves, going digital first. So actually there are pockets where it’s happening. And if I take one step back, actually the beauty of AI and healthcare, you can look to a lot of other segments in the world that are struggling to even find use cases for what AI is really going to contribute. We have tangible benefits that we can deliver. with AI. If I look to imaging, one of the spaces that’s the earliest, I think, spaces that this started to kind of really get an impact on, we have real tangible benefit. We developed one algorithm, SmartSpeed, that actually triples the productivity so you can do double the scanning in an hour that you could do before. It has massive adoption. One in four hospitals in Japan use it already now. So it actually is landing, is being adopted because it solves an immediate need and people can work with it. So there are examples where it’s happening. If you look to the accuracy level that we can get to now with the help of AI, if you look at how we can speed up. So I think that’s an area. But at the same time, if I look to pathology, one of the very important areas in healthcare to get to a complete patient diagnosis, only 10% is digitized. 90% is still the analog way. We have to do additional slides. They need to be kind of analyzed. There are not enough people to do it. So there still we are at the early stage of actually getting that into full digital kind of day and age. So I think there are two ways of looking at it. On a positive, we are moving. We have use cases. We have technology. And actually I think we are further ahead than some other segments that are still wrestling how to apply it and really get kind of the benefit from it. But to do this at scale is something that we still struggle with. And that’s where we need to have the system to play ball.

Karen Tso: Roy, thank you. And that dovetails neatly to Shobhana because the Indian market where you have a large footprint. Fascinating because in the private sector there are pockets of encouragement as Roy points out. So Apollo Hospitals is the largest profit hospital, for-profit hospital in India. You’ve been integrating digital and AI into your healthcare ecosystem across hospitals and improving access to telemedicine, for instance, some of the technology you have. Now, too often the public sector lags the private sector when it comes to digital. How does public health adopt some of the initiatives that you’re rolling out already throughout your health system?

Shobana Kamineni: Well, basically, you know, seeing that Gianrico and and Roy were here on this panel, I kind of threw out everything that I had prepared for this, in terms of what our hospital does, because it’s their hospital systems now, who can have the access, who have the people, the technology, the engineers and the finances can actually bring the best to bear for their patient. So let’s move that on one side and bring this conversation, like we do at Apollo all the time, we bring it back to the patient or we bring it back to the consumer. So how is AI really going to help you all? And then you can forget whether it’s the government that gives it to you or a private or whoever gives it to you, you will consume it in the way that you find most trustworthy because at some point people realize that they’ll have to pay for healthcare, whether you have some amount of insurance or you’ll find ways. So you do that or the country itself, we moved in India from 1% of spending on healthcare to 2% and it was a big shift for people to do for a developing country. Of course, the US it’s 15%. So governments really do spend money and I think what AI will help, if you look at it from the customer point of view, we just look at the customer. First, I believe that we are not the only providers for you all, for a customer. A customer will go to Google, will go to Amazon, will go to all these who are also healthcare providers, Apple, so your watch is probably your best. And these are the ones who are redefining what healthcare can mean for you because it has to be about you. It has to be able to predict earlier, it has to be able to give you the right nudges and these are all the language that we learn. ever since AI came into play with that. We had a different language of fear and urgency when it was COVID and then people came. So that was one that you know you need it when you need it. But there’s also this language now of how can I stay better? How can I stay well, live longer? And all this is what AI has been teaching you. And there’s where you’re consuming it and you’re going back to healthcare systems and saying, we don’t trust you. You’re not doing enough. You’re too expensive. And so we’ve come back and said, we got to solve for this. How can we scale? How can we make sure? So now I have 22 million people on my app with this and we’re sending them nudges with this. So, you know, because in India you can build beds but it won’t be enough. How do you keep 1.5 billion people, you know, healthy? And I think these are the conversations that we have fantastic partnerships. One is with WEF that we did this Edison Alliance and we just put, we used, you know, with American towers, we had digital dispensaries and we could map people’s healthcare that, you know, so women who didn’t want to go to a healthcare center were coming, you know, or digitally. So they were consuming. So this is a sector that is really left out in many countries, especially India. Another thing is, and I can go on, but let me close it by saying the cost of looking after them for preventive healthcare was $8 for us in India. It was $8 for the year for testing, for keeping their records and for giving them basic medication. So I think that this is a huge, you know, this is a solution for the disease burden of the world.

Karen Tso: So I’m bringing it back front and center to the patient. It’s an incredible number on the cost of healthcare. It’s fascinating. Nicola, let me turn to you. APATH is working in more than 70 countries to improve health, so a unique perspective when we talk about health equity and where to go from here. Now, tech has promised to pretty much democratise just about everything, but we haven’t really seen it yet when it comes to health equity. Realistically, how much impact do you think digital and AI will have when it comes to the next five years in achieving the sustainable development goals, at least for health, especially now that the US is out of the WHO as well? You were throwing that in in the end. Yes, I was. Someone was waiting for this. It was not in the script. I was afraid you were going to ask me about… Since I am from Denmark but live in the US, I thought you would ask me about Greenland, but you didn’t. Oh, I can’t ignore you, too. So, I’m happy to answer the other question.

Nikolaj Gilbert: No, so, I think we are way off track on the SDGs, that that’s not only the health SDG and that’s not very good for the world. I think without a digital transformation happening 10, 20 years ago, we would be even further back. So, I would say I would probably divide the answer into one AI and one on the digital transformation, excluding AI. So, I think many countries in the global south and the countries we work in has taken on a digitalisation of the public healthcare system. And, of course, as the backbone of their information systems for addressing epidemics and also… but also non-communicable diseases. So, that drive towards having a health system where you have the information that you need, a supply chain system with availability of data, of access to medicine and so on, is really something that countries have been able to integrate into their system. So, that has given them a good foundation to move forward and also deploying AI at a later stage. But, for me, digital is always… the perspective is as an enabler. It’s an enabler for health outcomes. So if you look, for example, health outcomes, also other areas, of course, but if you look at under five mortality rates, we know that vaccination is critical for young children. So how has digital immunization systems, patient records really helped increase the vaccination rates around the world and the countries we work in? So those are the kinds of investments in digital infrastructure and health infrastructure that has helped many, many countries. And I’m not seeing that going away in the next five years. I’m actually seeing more countries, and you mentioned India, telemedicine to access for populations in rural areas that doesn’t have access to health clinics, health facilities. So big opportunities there to roll out telemedicine further over the next couple of years. So I think that still is giving us some promise that we will reach more people by using digital systems that enables us to reach those people. It has to be combined with access to medicines, affordability of those medicines, access to doctors, nurses, community health workers that needs to be equipped to actually deal with those patients, but also being able to utilize those systems, which is gonna be essential. And we need also to continue the political will. So every government around the world needs to take on digital as an enabler for a strengthened public health system in the country. Without that political commitment, it’s very hard to roll out public sector reform, including establishing a digital infrastructure. We’ve seen this in the G20 in India, for example, which had digital as one of the key priorities, Brazil, South Africa, next this year, they’re gonna talk about universal health coverage, primary health care is gonna be a priority and primary health care, universal health coverage needs digital infrastructure to work. So I say we’re gonna see some influence on the. the SDGs on the digital transformation. On the AI, I think for the vast majority of the populations around the world, which lives in the global South, it’s not gonna move the needle a lot. We are currently deploying AI technology in three countries in Africa in a project with the Gates Foundation. We are very grateful for that partnership. It’s the first test of how AI can be used in clinical decision-making by community health workers in rural areas. So we’re gonna apply AI learning and see if they can actually use that when they see patients integrating it with the regulatory system and so forth in the country. So we’re very excited with this pilot, but it’s not gonna deliver any major results in the next couple of years. I don’t wanna go too far off script, but I do want to ask the other panelists

Karen Tso: about the US stepping out of the WHO and whether you see an impact here along with the massive investments we’ve just heard this week, $500 billion going into AI investment in the United States. As we talk about digital and AI being an enabler, the infrastructure will be very key here and what impact it has. Shobhana, do you wanna touch on that? What the US is doing, stepping out of the WHO, but building its infrastructure and technology. Does that have an impact leaving a big divide between say the United States and India?

Shobana Kamineni: I hope there’s no pandemic in the next four years or any critical major event. But having said that, I think that WHO does play an important role in unifying, but I don’t think it is the only thing that, it’s not gonna take down, make or break India’s healthcare system, frankly. So it helps to be, we like to be part of bigger networks. It helps in that aspect. I don’t think that we’re at a stage that we depend on them.

Karen Tso: Roy.

Roy Jakobs: If you think about what I just said, that we need the systems approach. Actually, I think we will miss certain support and oversight that a more neutral body like the WHO can provide to kind of cross learning. The other one, and we discussed it today, if you talk about systems approach, you need to have public-private partnership, but actually there’s also a lot of public-private and philanthropic or kind of partnership that is going into healthcare, and we need to orchestrate that. Again, WHO has played an important role, including sustainable financing, because financing is a big part of healthcare, and especially if you talk about health equity, I think WHO has played an important role as well in terms of making sure. So I think we will miss out if WHO gets stronger. I also agree it’s not the only role or it’s the only actor, so in that sense we will need to, as a system, pick it up, but it’s something that I think we will miss if they would be weaker.

Karen Tso: John Rieke?

Gianrico Farrugia: There isn’t much to add there, because this is fresh, we’ll have to see how it unfolds. What is clear is that other organizations will have to step up and will have to create new partnerships, and sometimes that does give us an opportunity to do something a little differently from the past. And the fast-moving space that is the digital transformation of healthcare, that is not

Karen Tso: necessarily a bad thing. Thank you all for taking that question. I want to talk more about enablers, and Roy, if I can come to you. We know it’s not enough just to have good technology. If you lack the capital, you lack the incentive scheme, it’s skewed towards short-term profit, for instance, switching costs, and of course we saw that with even just digital health records, there was switching nightmare, because a lot of medical systems just didn’t want to pay for that upfront cost of switching medical records online. In terms of what we see here, too much regulation, you pointed out as well, what are we chasing in terms of systemic impact on health? Now, Philips has been involved in various initiatives here, tackling value-based healthcare, health systems, sustainable resilience. What is the most important public and private enablers that we need to focus on?

Roy Jakobs: to focus on globally. What can you tell us about these enablers? I think, and I said it before, if you want to really leverage the power of AI, you need to free the data first and actually make the data flow. So no silos. And do it around the patient. So the patient data need to be put in the hands of the patient first or put in the hands of kind of the body’s governments that actually can orchestrate it in the right way. Currently, data are locked up in silos across the world still, that actually makes it, you cannot travel. If you’re a patient, it’s very hard for you to travel with the data or to kind of keep them. And if you want to enrich them and put the AI on top, I think, so that’s for me, the data. And the data challenge is one big one, big enabler to unlock power of AI. The other is kind of how we train the algorithms with the right clinical data. As you know, the common algorithms, they leverage the web to get their knowledge from. We need actually to train it on proprietary data. And also there, if you look kind of, what are the data sets that we can train them? We know they will get more powerful if they get more data. Currently, you’re still limited in some of the data sets that we can use to train them. So I think the better we can train them, expose them to kind of the right data sets and also integrate those data sets. Because I was talking about imaging pathology together with genomics data together, that actually gives you a really full view of the patient and the situation. Now, there are not a lot of kind of current bodies that I can see that can do that in an easy and digital manner. So how can we kind of really accelerate that journey? And there’s a lot tangibly that is being done that you can do. So that I think is another kind of critical enabler. And then I would still, next to the patient, I would also still put the user, the provider at the center. Because if the nurse doesn’t want to kind of adopt it because they’re afraid of it, or it’s not user-friendly, they will, we will not get the benefit. If the doctor thinks that he’s going to be replaced by it, which he should not, because I think it will actually enhance, they will reject it. And we were also discussing early today. We have multiple generations of doctors and nurses that actually we need to deal with. So how do you actually get them along in the journey, educate them, train them, bring them – kind of bring them to ease with it? I think one of the big changes or I think the benefits that healthcare has, I think in healthcare you should not be – nobody should be afraid that AI is going to take their job because we don’t have jobs enough to be filled with people in healthcare. The biggest kind of problem for healthcare is that we don’t have kind of people to take care of the people. So actually healthcare has to come to the rescue to actually fill the void that actually there is. And let’s do it on the tasks that actually they don’t want to do, the routine tasks. If a nurse spends 15 minutes an hour on admin tasks, let’s take them away. If a doctor has to do prescription and kind of a lot of reporting, let’s actually go there first. Those are routine tasks that we can actually really go after hard with AI. And that’s a quick win. It’s also a lower risk. The closer you get to precise medicine, diagnosis, treatment, intervention, kind of AI guided robotics that kind of take over, then you get into pretty exciting areas, which I think will come. But we better train them well, we better test it well, and we know what we do. So I think there’s a lot of application area, but let’s start with the basics.

Karen Tso: You need to have the data set, you need to have the right, properly trained algorithm, and you need to have the patient sees the benefit and the providers that know how to use it. I think that’s a fascinating point about the data. It takes me back to a conversation I had with Tim Berners-Lee about who owns what when it comes to data. And of course, a very basic example being social media, that if you’re with one platform, you can’t transport your data elsewhere. He’s kicking along the next level when it comes to all data, not even healthcare. Perhaps that data can be transported, no matter what it is, whether it’s driver’s licenses, whether it’s healthcare identity. That is a fascinating point that data could be transported by the patient. Does anyone else want to come in on that? So I kind of agree with you, but I think that just giving data freely to you guys, unless you’re in partnership with us.

Shobana Kamineni: So there has to be constructs of good partnerships. I mean, we’ve done some great things with you all. We can build, you can build better algorithms when you have the right, when you’re working as partnerships. If you’re just going to let all your data out, you don’t know what kind of, what will happen with it, the hallucinations, all that, you know, there’s patient safety, there’s this thing that data belongs to the patient. And so all those, you know, we come up with ethical issues. So it’s not that simple that you can just let data flow freely. This will be regulated. It’ll always be a challenge. But I think what we now see is really good partnerships because, you know, just imagine that we did a project with Google, just like you did, you know, what you had explained that you can, that they can actually look and see, you know, from, they can see if a patient has TB. Way ahead, using digital tools. So then you can take it out to large population because so many people in India are vulnerable to tuberculosis, and that’s dangerous, you know, because of the antibiotic usage. So for that aspect, these are the things that you can come and collaborate on. So yes, I do think that the data is important,

Karen Tso: should flow, but not that freely. John Rico, let me turn to you around responsible technology here at AI and Trust. These are big issues, particularly when it comes to the healthcare system. Nobody wants bias in the system. And of course, nobody wants any potential harms to come to patients by unlocking some of this data. How do we strike the right balance here between innovation and responsibility, particularly in the intelligent age of AI?

Gianrico Farrugia: So we clearly are not there yet, and we should be clear-eyed about why we’re not there yet. And I would argue that healthcare has some responsibility. responsibility, governments have some responsibility, we all collectively have, is that we seem to be in a position where everybody wants to dip their toe into the water, but unless the temperature is exactly right, they don’t want to jump in. And that is not good because healthcare at the moment is not in a good place. And therefore, even if the temperature isn’t exactly right, there still is a tremendous amount of benefit from getting in and then making it better. I personally would not want to have my healthcare in some specialties without AI, because I firmly believe that I will get a better outcome. So we need trust. We have to do a better job defining what that means. Inadvertently, we created the impression that trust means that everybody who uses AI has to understand exactly what’s going on, no black boxes. Well, that’s not realistic. I do not know of a single x-ray technician that uses one of Roy’s machines that understands truly what goes on in the machine, but Philips plus some sort of authority have said it’s okay to use and what it tells you is actually what is going on. I’m a big proponent for assurance labs, private public partnerships that come together that allow anybody to take their algorithm and be able to say that it’s fit for use and it does what it says it’s going to do. In our platform, we call that validate. We recently had an example of a company that developed an algorithm to diagnose epilepsy at home. Epilepsy affects about 1% of the population. It’s a big deal. They ran their algorithm to our platform and we found it did well overall, but there was a subset, one subset in the public that didn’t do as well. They had to go back, arrange that and they now have FDA approval. By pushing the responsibility To independent groups that then are able to then verify the algorithm But more importantly come back to it over time. These are self-learning algorithms You can’t lock them down so you can lock them down and you have to come back to them You have to come back to them. You have to come back to them if you do that I truly believe we will start getting back the trust that we need In order to get back to what we’re all talking about the patient It’s not in the patient’s best interest to not trust unfortunately, they should not trust at the moment But it’s up to health care to install that trust So we have to stand up to do it and one way to do it is to make sure that Providers also trust because at the moment providers don’t trust either and I am a firm believer that the way to do that is to Decentralize AI AI needs to be not run from a centralized place It has to be done run within clinical departments within units in in companies because you will learn more and you’ll be able to learn about AI from your co-worker from the person next to you in the cubicle from The physician that is talking to the patient in the bed And that’s where I think you start building those elements of trust in the in the health care system That then also translates into a patient rest

Karen Tso: Nikolai do you want to come in on this because the different countries you work around of course different levels of trust and we know that This is very personal aspect of our lives to be trusting people with our data To manipulate it to crunch it to try and target us with solutions. How do you think the approach should be here?

Nikolaj Gilbert: It’s interesting because the u.s. Of course has the main technology companies that are driving this technology But I’m actually not sure that u.s. Is gonna be number one in terms of applying the technology because of the risk in the u.s. In terms of lawsuits and so forth which Means that that you and also the way the health system is set up There are other trust based more trust based societies where I think there is a better trust between providers Physicians and patients and Maybe I would start in looking at those societies where you can actually where you already have a strong established trust between the system and its population and That so people in the u.s. Don’t go off into the doctor. Of course And we know that because of finance and because of lack of trust and so on. So I do believe that the deployment is going to happen maybe somewhere else where the trust is different. I don’t want it to happen in a country in Africa where there is no guardrails, of course, because that’s where it can go really, really wrong. So that’s why I talked about this project we’re doing now to see under which conditions can this actually work, under which regulatory environment, under which health education program can this actually work, and how do you build that trust among the patients and the community where you are delivering care. So it needs to deploy in a place where there is already trust for it to show the results that we want to show that this can actually help address some of these workers’ shortages that you mentioned, which I truly believe in, that it’s not going to replace work. It’s actually going to solve some of the gaps we have in the countries that we work in, in terms of providing health care advice, education and so forth.

Karen Tso: My mind went so many places then when you were speaking, one about the US being in the fast lane when it comes to AI, but not when it comes to health and AI.

Nikolaj Gilbert: That was correct. Yeah, that’s what I think.

Karen Tso: And the other point that if the continent of Africa is getting access to technology perhaps at a subsidized price, then maybe state actors are involved in extracting the data from some of these core systems. Yeah. So there are many complexities in this, and we have to know what we’re going into, and

Nikolaj Gilbert: we have to know the environment and the country that we work in, because it can go wrong.

Karen Tso: So yeah, that’s my point. Roy? Yeah, I made a few comments. So first, and I think on the data liquidity, what I was trying to say is kind of first we need to make data digital. So In Germany, 10% of data is still fax. Hard to do. Digital pathology, 10%, yeah, it’s still fax. So it’s hard to do AI on fax, right?

Roy Jakobs: It’s, if you have pathology data, 90% is analog, right? You cannot do AI. That was my point on data. Secondly, we have a responsibility, right? Federico and myself have been part of the National Academy of Medicine for self-conducting mechanism on responsible AI. So I think we also need to be able to look into the mirror every day to say, okay, are we doing the most? On the other side, we also need to be fair to AI. We should not ask more from AI than we ask from a doctor. And we do. On the accuracy level that we test, we actually put levels in of accuracy to AI that actually doctors do not meet. The consistency of what AI can deliver versus what the consistency is that if you go to certain doctors’ practices, it’s not always there. So I think we should be at least where doctors are better, but we should also not say you need to do the kind of impossible because you would wipe out all the benefits and take the patient perspective. I can tell you, if you’re that late stage cancer patient fighting for the ultimate resolution that’s not yet there, you will want AI to explore the world for the last resort that you are craving for to help you survive. So I think there are different perspectives in terms of, and that is kind of what fairness is about in AI. I think it’s extremely important we do it right, we get it right, but we should not curve the potential of this potential revolution in technology and how it really can progress healthcare and the outcomes of healthcare. And still, we need to do it in a responsible way.

Karen Tso: True. Shobana, let me come to you on some solutions, avenues here for possible solutions, international collaboration, partnerships between healthcare and tech. What are some of the solutions you see here? So one is that I agree with Roy that these are the collaborations that will bring together the best because it can’t be done just with the hospital system.

Shobana Kamineni: But when you collaborate with whether it’s, we just opened in the state of. of Andhra Pradesh in my father’s village. We have a medical college and the University of Leicester just came day before yesterday and we have a center of innovation with this. So I think that’s something that AI can do that it’s not about a geography. You can put it there, you can get these collaborations together and these partnerships don’t need to be this high-end exclusive in bubbles, but we need to put more of them around the world. And the more that we have, the more that people can, that you will get those. And I agree that you can’t expect more than what a doctor does, but a better perspective for that, I mean, actually a conjunct to what you said is that the reason that we did our clinical intelligence engine is because when you go to doctors, each one will have a different interpretation sometimes and to get there. So what we can do is that 80% of it, if AI can actually smoothen the process for the patient, it actually makes their life so much easier, understandable, and then that 20% has to be done by a doctor. The validation has to be done by a doctor. A, it makes him more productive, he can do more cases and the cost comes down for that particular doctor visit. And the third part of it is that it just gives a much better experience because then everything is there. That instead of a doctor, the 15 minutes he would have spent earlier, this actually makes it so much better. So.

Karen Tso: Shobana, thank you very much. We are running out of time. So I just wanna whip around the panel and get some action here, some takeaway. If I can start with you, Nikolai, based on everything we’ve just discussed here, if there’s one thing that you feel should be done between public and private collaboration in the next 12 months to advance what we’re talking about here, digital health on a global or regional scale, what would it be? 30 to 40 seconds.

Nikolaj Gilbert: Yeah, I think standardization, interoperability, applications that will be easy to integrate into public systems, because innovation is very hard to take into existing health information systems, so anything that private sector develops needs to be easily integrated for it to

Karen Tso: be really usable for integration into the system and actually benefit the population. Gianrico, same question to you. If there’s one thing you feel could be done via public-private collaboration over the next 12 months, what would it be?

Gianrico Farrugia: We need to move healthcare from a pipeline to a platform model. We need to do that by creating defined pathways that are coupled to financing models that allow countries, wherever they are, to know what the next step is, and then to make the choices that allow then patients to have better care, deciding to either partner with somebody else or solution developers to develop solutions or develop their own solutions. If we do that and tackle that underlying problem, then not only will patients benefit, but the system itself will continue to self-learn and create truly better outcomes for all.

Shobana Kamineni: Shobana. I’d go all in in reskilling, upskilling and teaching with this in the next 12 months if I have that, because that wave is going to come for us to take, to seize that moment. We need to have multiple people who can actually take charge of this, so skilling, reskilling, super important.

Karen Tso: Roy. I think from the adoption, because it’s all about how we drive adoption, and I

Roy Jakobs: think rallying up the financing with regulatory support and how we kind of have the support to the practitioners, I think will define how AI will find its way into the daily practice. You’re also succinct, which is amazing. We actually

Audience: have time, we might have time for a very quick question. Does anyone have a… here we go. Stand up. Okay, sure. Sorry, I actually in the meantime ask AI the question which country is really collecting all available health data in the central system and The judge PT answer is like, okay, Switzerland is the only one I found But many people many opinion claims that actually they claim to have this data, but nobody have seen this data So I will ask you the same question. So Have you ever come across? The country which really collects everything including, you know, I mean rankings Resonance, etc. All the health data of the people in the central system Can anyone name one I Don’t I I understand the question. I would say that’s not the important question. The important question is Data in a position that can be used so there’s many there are several countries that are collecting data They’re putting it in a in a in containers that are not conducive to digital first

Gianrico Farrugia: There there’s I will not name particular countries because of what I just said, but I let others chime in boy ten seconds Yeah, so maybe one great example actually is Indonesia

Roy Jakobs: So Indonesia what what they did actually they had the kovat app They kept it alive where they actually, you know, I’m covered every everyone had their own covert app proving that kind of day We’re vaccinated or not They use that and by now they kept that they actually have it for every inhabitant of Indonesia and every single test Medical procedure is loaded on that app, right? So there they have it on that app in one in one point of Nigga you wanted ten seconds. No, I just say I

Nikolaj Gilbert: Think it’s a good question. But the question it’s You don’t need to have all the data to actually deploy digital Technologies and do a digital transformation and utilizing AI so in the perfect world you will collect all these data, but 90%, 80% would also give you a good start, as long as you focus in the areas where you collect the data. Shobana, you were whispering China?

Karen Tso: No, I like all, everybody’s, and we’re out of time. We are out of time. Thank you so much for your participation today, and thank you to our panelists. You were great, including on timekeeping. Thank you. Thank you. Thank you.

Agreement Points

AI has potential to improve healthcare outcomes

speakers

– Roy Jakobs
– Gianrico Farrugia
– Shobana Kamineni
– Nikolaj Gilbert

arguments

AI can enhance productivity of healthcare workers

AI can improve diagnosis, treatment and prevention of diseases

Digital tools enable low-cost preventive care at scale

AI can help address healthcare worker shortages

summary

All speakers agree that AI and digital health technologies have significant potential to improve healthcare outcomes, increase efficiency, and address workforce shortages.

Current healthcare systems are not optimized for AI integration

speakers

– Roy Jakobs
– Gianrico Farrugia

arguments

Healthcare systems struggling to fully leverage AI despite its potential

Antiquated healthcare architecture not suited for AI integration

summary

Both speakers highlight that existing healthcare systems are not well-suited for integrating AI technologies, which is hindering the full realization of AI’s potential in healthcare.

Data liquidity and interoperability are crucial for AI adoption

speakers

– Roy Jakobs
– Nikolaj Gilbert

arguments

Lack of data liquidity and interoperability between systems

Improve standardization and interoperability of health IT systems

summary

Both speakers emphasize the importance of data liquidity and interoperability between healthcare systems for effective AI implementation.

Similar Viewpoints

Both speakers stress the importance of properly training and validating AI algorithms using clinical data and through partnerships between public and private sectors.

speakers

– Roy Jakobs
– Gianrico Farrugia

arguments

Need for better training of AI algorithms on clinical data

Develop public-private partnerships to validate AI algorithms

Both speakers highlight the potential of digital health technologies to improve healthcare access in developing countries and rural areas.

speakers

– Shobana Kamineni
– Nikolaj Gilbert

arguments

Digital health can increase access to care, especially in rural areas

Potential for digital health to advance SDGs in developing countries

Unexpected Consensus

US lagging in healthcare AI applications despite technological leadership

speakers

– Roy Jakobs
– Nikolaj Gilbert

arguments

Only 10% of pathology is digitized, limiting AI applications

US leading in AI technology but lagging in healthcare applications

explanation

Despite the US being a leader in AI technology development, both speakers unexpectedly agree that the country is lagging in applying these technologies to healthcare, which is surprising given the US’s reputation for medical innovation.

Overall Assessment

Summary

The speakers generally agree on the potential benefits of AI and digital health technologies in improving healthcare outcomes, increasing efficiency, and addressing workforce shortages. They also concur on the challenges faced in integrating these technologies into existing healthcare systems, particularly regarding data liquidity, interoperability, and the need for proper training and validation of AI algorithms.

Consensus level

There is a high level of consensus among the speakers on the potential of AI and digital health technologies, as well as the challenges in their implementation. This consensus suggests a shared understanding of the current state of digital health adoption and the steps needed to advance it. However, there are some differences in focus, with some speakers emphasizing global health perspectives and others concentrating on specific technological aspects. This diversity of perspectives provides a comprehensive view of the challenges and opportunities in digital health adoption across different contexts.

Different Viewpoints

Data sharing and privacy in healthcare

speakers

– Roy Jakobs
– Shobana Kamineni

arguments

Lack of data liquidity and interoperability between systems

Digital tools enable low-cost preventive care at scale

summary

Roy Jakobs advocates for greater data liquidity and interoperability, while Shobana Kamineni emphasizes the need for careful data sharing through partnerships to protect patient privacy and safety.

Approach to AI adoption in healthcare

speakers

– Gianrico Farrugia
– Nikolaj Gilbert

arguments

Create new healthcare architecture suited for AI integration

Need to consider country-specific contexts for AI deployment

summary

Gianrico Farrugia proposes a new healthcare architecture for AI integration, while Nikolaj Gilbert emphasizes the importance of considering local contexts and existing trust in healthcare systems for AI deployment.

Unexpected Differences

US leadership in AI healthcare applications

speakers

– Nikolaj Gilbert
– Other speakers

arguments

US leading in AI technology but lagging in healthcare applications

explanation

Nikolaj Gilbert’s observation that the US might not be the leader in applying AI to healthcare, despite being at the forefront of AI technology development, is unexpected given the general perception of US healthcare technology leadership.

Overall Assessment

summary

The main areas of disagreement revolve around data sharing, privacy, and the approach to AI adoption in healthcare across different contexts.

difference_level

The level of disagreement among speakers is moderate. While there is general consensus on the potential benefits of AI and digital health, there are significant differences in how to implement these technologies, particularly considering privacy concerns, regulatory environments, and local healthcare system contexts. These differences highlight the complexity of implementing digital health solutions globally and suggest that a one-size-fits-all approach may not be effective.

Partial Agreements

All speakers agree on the potential benefits of AI in healthcare, but they differ in their approaches to implementation. Roy Jakobs focuses on enhancing worker productivity, Gianrico Farrugia emphasizes improving medical outcomes, and Shobana Kamineni highlights increasing access to care.

speakers

– Roy Jakobs
– Gianrico Farrugia
– Shobana Kamineni

arguments

AI can enhance productivity of healthcare workers

AI can improve diagnosis, treatment and prevention of diseases

Digital health can increase access to care, especially in rural areas

Similar Viewpoints

Both speakers stress the importance of properly training and validating AI algorithms using clinical data and through partnerships between public and private sectors.

speakers

– Roy Jakobs
– Gianrico Farrugia

arguments

Need for better training of AI algorithms on clinical data

Develop public-private partnerships to validate AI algorithms

Both speakers highlight the potential of digital health technologies to improve healthcare access in developing countries and rural areas.

speakers

– Shobana Kamineni
– Nikolaj Gilbert

arguments

Digital health can increase access to care, especially in rural areas

Potential for digital health to advance SDGs in developing countries

Key Takeaways

Digital health and AI adoption in healthcare is happening in pockets but not yet at scale globally

Major barriers to adoption include antiquated healthcare architecture, lack of data interoperability, and trust issues

AI and digital health have significant potential to improve healthcare access, quality, and efficiency

Public-private partnerships and system-wide approaches are needed to advance digital health adoption

Different countries are at varying stages of digital health implementation, with some emerging markets moving quickly

Resolutions and Action Items

Create new healthcare architecture better suited for AI integration

Develop public-private partnerships to validate AI algorithms

Focus on building trust through responsible AI development

Upskill and reskill healthcare workforce on digital technologies

Improve standardization and interoperability of health IT systems

Unresolved Issues

How to achieve data liquidity while maintaining patient privacy and security

Best approaches for integrating AI into clinical workflows and decision-making

Regulatory frameworks needed to govern healthcare AI across different countries

How to ensure equitable access to digital health technologies globally

Long-term impacts of AI on healthcare workforce and job roles

Suggested Compromises

Balance between data sharing for AI development and protecting patient privacy

Set realistic accuracy standards for AI in healthcare, comparable to human clinicians

Combine AI capabilities with human expertise rather than fully automating healthcare

Develop AI solutions that can be easily integrated into existing health systems

We know that healthcare plus AI is better than healthcare on its own. We know that you can diagnose, treat, and even prevent disease in ways you couldn’t do three years ago. So why the gap? It’s because at the moment, healthcare organizations, governments are unable to take what’s currently available and meet the needs of what populations expect from them. And I believe that gap is structural. That gap is because we’re trying to overlay AI and digital healthcare onto an architecture for healthcare that is antiquated and doesn’t work.

speaker

Gianrico Farrugia

reason

This comment insightfully identifies the core challenge in implementing AI in healthcare – the mismatch between new technologies and outdated healthcare infrastructure.

impact

It shifted the conversation from discussing the potential of AI to examining the systemic barriers to its adoption, prompting others to consider structural issues in healthcare.

So how is AI really going to help you all? And then you can forget whether it’s the government that gives it to you or a private or whoever gives it to you, you will consume it in the way that you find most trustworthy because at some point people realize that they’ll have to pay for healthcare, whether you have some amount of insurance or you’ll find ways.

speaker

Shobana Kamineni

reason

This comment reframes the discussion from an institutional perspective to a patient-centric view, highlighting the importance of trust and accessibility in AI adoption.

impact

It broadened the scope of the conversation to include consumer behavior and trust, leading to discussions about how to make AI solutions more accessible and trustworthy for patients.

I think without a digital transformation happening 10, 20 years ago, we would be even further back. So, I would say I would probably divide the answer into one AI and one on the digital transformation, excluding AI.

speaker

Nikolaj Gilbert

reason

This comment provides important context by distinguishing between general digital transformation and AI-specific advancements, offering a more nuanced view of progress in healthcare technology.

impact

It led to a more detailed discussion about the different stages of technological adoption in healthcare, from basic digitalization to advanced AI applications.

I personally would not want to have my healthcare in some specialties without AI, because I firmly believe that I will get a better outcome. So we need trust. We have to do a better job defining what that means.

speaker

Gianrico Farrugia

reason

This comment boldly asserts the value of AI in healthcare while acknowledging the critical need for trust, highlighting the tension between innovation and public confidence.

impact

It sparked a discussion about the balance between pushing for AI adoption and ensuring public trust, leading to conversations about responsible AI development and deployment.

We should not ask more from AI than we ask from a doctor. And we do. On the accuracy level that we test, we actually put levels in of accuracy to AI that actually doctors do not meet.

speaker

Roy Jakobs

reason

This comment challenges the common perception of AI in healthcare by pointing out the double standards applied to AI versus human doctors, offering a fresh perspective on AI evaluation.

impact

It prompted a reconsideration of how AI in healthcare is evaluated and judged, leading to discussions about fair standards for AI performance and adoption.

Overall Assessment

These key comments shaped the discussion by shifting focus from the potential of AI in healthcare to the practical challenges of implementation, including structural barriers, patient trust, and fair evaluation standards. They broadened the conversation from a purely technological perspective to include systemic, ethical, and patient-centric considerations, ultimately painting a more complex and nuanced picture of the role of AI in transforming healthcare globally.

How can we create a new architecture for healthcare that better integrates AI and digital solutions?

speaker

Gianrico Farrugia

explanation

Farrugia suggests that the current healthcare architecture is antiquated and unable to fully leverage AI and digital solutions. Creating a new architecture could enable better scaling and implementation of these technologies.

How can we improve data liquidity and sharing in healthcare while maintaining patient privacy and trust?

speaker

Roy Jakobs

explanation

Jakobs emphasizes the need for better data flow and integration to fully leverage AI in healthcare. This is crucial for training algorithms and improving patient care across different healthcare providers and systems.

What are effective ways to build trust in AI and digital health solutions among both healthcare providers and patients?

speaker

Gianrico Farrugia

explanation

Farrugia highlights the importance of trust in adopting AI solutions and suggests exploring methods like assurance labs and decentralized AI to build this trust.

How can we accelerate the digitization of healthcare data, particularly in areas like pathology?

speaker

Roy Jakobs

explanation

Jakobs points out that a significant portion of healthcare data is still analog, which hinders the application of AI. Accelerating digitization is crucial for advancing AI in healthcare.

What strategies can be employed to effectively reskill and upskill healthcare workers to use AI and digital health technologies?

speaker

Shobana Kamineni

explanation

Kamineni emphasizes the importance of training healthcare workers to use new technologies, which is crucial for the successful implementation of AI and digital solutions in healthcare.

How can we develop standardized, interoperable AI applications that can be easily integrated into existing public health systems?

speaker

Nikolaj Gilbert

explanation

Gilbert stresses the need for AI applications that can be easily integrated into existing health information systems, which is crucial for widespread adoption and benefit to populations.

What financing models and pathways can be created to help countries at different stages of development implement AI and digital health solutions?

speaker

Gianrico Farrugia

explanation

Farrugia suggests the need for defined pathways coupled with financing models to guide countries in implementing AI and digital health solutions, regardless of their current level of development.