 Welcome. I'm Jennifer Schenker, editor-in-chief of The Innovator, a global publication about digital transformation. A few housekeeping points. The session is being publicly live-streamed and will be available on the forum's web pages. Please use the Zoom chat feature if you want to raise questions. I'll do my best to get to them. The 30-minute panel will be followed by a more detailed discussion that will be limited to forum members and partners. Today's discussion will focus on the new healthcare environment with data-driven AI. Artificial intelligence has the potential to speed up drug development and our ability to eradicate diseases. It can help get medical advice to people who have no access to doctors and usher in personalized medicine, making more precise diagnosis or prediction of diseases and response to therapy possible. The more data AI collects, the more accurate it will be. But what happens if we give away all the data and leave all of these wonderful AI-led discoveries to a small number of global companies? Will we end up worse off with even more unequal access to medical care? Designing an approach to ethical and equitable data-driven AI in the healthcare sector raises many questions. Should we allow every single molecule in people's bodies to become exploitable for economic value? Or should health data be considered a public good and treated separately from data about which brands we prefer or what we're posting on social media? Who owns an individual's health data and how do we make certain that privacy is respected? How do we guarantee that individuals and or communities benefit from the data collected about them? And how do we ensure that data collected represents and is designed to help members of all races and ethnic groups and that low-income countries or communities do not get left behind? I'm pleased to introduce an all-star panel to address these crucial issues. The Honorable Stella Nedaveni-Abraham's South Africa's Minister of Communications and Digital Technologies, Dr. Gianrico Farugia, President and CEO of the Mayo Clinic, and Carlos Correa, Executive Director of the South Center. I'd like to start with you, Minister. Sub-Saharitan Africa has become an emerging arena for digital health innovations directed at strengthening healthcare in terms of both patient management and disease surveillance and prevention. Across the region, with the world's largest burden of disease and the most severe shortage of healthcare workers, we see increased use of digital health solutions and interventions for clients, healthcare providers, health system managers and data services. For example, the Rwandan government has teamed with UK Startup to roll out an AI-powered digital diagnostic service with the support of the Bill and Melinda Gates Foundation. But in most parts of Africa, telemedicine is currently limited by the availability of basic infrastructure, such as study, electrical power, cellular network coverage, and broadband internet service. Please tell us how South Africa is tackling the infrastructure issues. Thank you so much, Jennifer, and greetings to all Japanese and everybody who's listening to us here. You are raising a very critical issue that actually, by the emergence of COVID, really exposed many governments in the realities that we have taken for granted. You are raising an important issue when you make mention of the AI, especially in that that one to achieve as we deal with this horrific virus that you are looking at. But as you correctly put it, the baseline of everything before we talk about deploying all those technologies, be it AI or any other, people have to be connected. Now in a country just like South Africa that has lots of digital divide, it becomes imperative that all governments prioritize that. The president of South Africa with his government, they've made sure that they establish the infrastructure council that looks into ensuring that the entire ecosystem of the infrastructure is addressed. And that's not only looking at the digital infrastructure, although we are saying, yes, the digital infrastructure will be the key driver or enabler of the services that must be rendered. We're looking at the infrastructure for the health sector. We're looking at the infrastructure for human settlement because as people are displaced, again, there's the super spread activities that we get to see. But at the center of that is that often showing that South Africa has a clear plan towards leveraging on the fourth industrial revolution that we're talking about. And at the center of the fourth industrial revolution that we're talking about, artificial intelligence is therefore there, which is why when we establish the presidential commission, we then took a clear resolution that there's a need for us to establish an AI center in order to make sure that we build that capability as we want to deploy AI in different aspects in the health sector, in the education sector, as I spoke about the digital divide, but also the limited resources that we have. These are the interventions that we're bringing there in order to be able to connect the citizens that are in the remote areas with specialists in certain areas. And as I'm speaking to you, the Department of Health is busy right now rolling out telemedicine to ensure that our people can be able to have that, but we've got to address basic things. The last time we came to Davos, Jennifer, they mentioned the issue of our load shedding. What is South Africa doing to make sure that we stabilize energy? Because we may be talking about the digital infrastructure, but if we're not looking into the entire ecosystem, then our people will not be able to enjoy the benefits that they should be enjoying in the technologies that we're talking about, which is why we're addressing the energy crisis that we have in the country. You've heard our commitment as a government, and of course at the center of that is looking into the transport issues, and I can go on and on. Our department is championing the rollout of broadband connectivity, which once again has been focusing on connecting public buildings and COVID has shown us that people may not have access to public buildings. They must therefore have access to connectivity in their respective areas, and that's what we're busy with now. Thank you so much, Minister. So if I can ask us a follow up, even if infrastructure issues can be overcome, some people question whether it's wise to allow AI powered health bots to diagnose people at a distance. What safeguards do you think should be put in place before allowing AI to do triage? And in your opinion, do we still need humans in the loop? Yes, I'm going to start with the last one. We do need humans in the loop. But as we talk about the basics before we deploy AI, at the center of this is to look into the ethics of AI, therefore looking at the programmers themselves. What are these things that they should comply with? Because if we fail to do that, we'll experience a situation whereby we have people who are discriminated against, and also people cannot attend or get access to certain services, which is why in South Africa we said the ethics of AI is something that's very important. That's why we are working on our AI strategy to say, how can South Africans make sure that as we deploy these technologies, they can have trust in them and we can increase public confidence so that as they continue to consume where they're consuming the services, but they're also guaranteed that they will not then face any talent that will be discriminated against them or anything like we have seen with the images of the robotics and all that, teaching them the ethics and the culture of the people of South Africa in order to make sure that it adapts to the African character that we're building. Thank you so much. So let me just ask you another question. Should health data be considered a public good and treated differently from other types of data collected? I want to use a concrete example here about who owns the data. The Israeli government recently made a deal with Pfizer and in return for access to vaccines that agreed to supply citizen data to the pharma company. In press interviews, an expert in digital privacy at the Israel Democracy Institute questioned the ethics of a deal that could bring millions in profits to Pfizer. She also said sharing large quantities of information could still potentially put individuals' privacy at risk, even if it's supposedly being made anonymous. Would the South African government be willing to trade its citizen's data for access to vaccines or medicines? If yes, what kind of safeguards would need to be put in place? And if not, why not? Well, then if I let me tell you that that question is a tricky one because I don't believe that it's a matter of yes or no. Because as you asked about the basis that must be laid by anybody who wants to deploy those technologies, this also relies on the complexity of ensuring that those rights, individual rights that we're talking about, they are protected. But again, as much as we recognize the importance of protecting individual rights, we also understand that those rights are not absolute. And therefore, in the event, for example, as you are faced with COVID, when we say we want to make sure that we can have minimized the spread of the virus, we can deploy such. But then how do we then control the people that must have access to it, which is something, for example, here at home when we introduce the track and trace, we said the Department of Health must take charge of that because specifically we wanted the data that will be gathered to go and help to plan better and make sure that people indeed are restricted and alerted on time. But at the center of that, it is a constitution that enshrines the individual rights. But as I said, there is the right to life. Therefore, if the measures that we're introducing seek to protect people's lives, therefore, there is nothing wrong. We cannot therefore say government must hide data because this data, as we understand the importance and the fact that it is a driving force in terms of the fourth industrial revolution era that we're participating in, we also understand that it must be availed in order for government to plan better, also for private sector and small businesses, so that as we talk about a potential, for example, potential, another outbreak that we may have, people know exactly what is happening and therefore the scientists can plan properly. But if we're going to want to monopolize access to data, we're not going to be achieving everything that we should be achieving in the ecosystem. So, yes, we do understand that data needs to be protected, but not at the expense of the country's livelihood and everything that goes with it. Thank you so much, Minister. I now like to turn to Dr. Farugia, the president and CEO of the Mayo Clinic. Data welcome and data is health data issues are very complex. What is your view on whether AI-derived health data should be considered a public good and treated differently from other types of data? And if so, what would that look like in practice? Should life-saving knowledge derived from AI be exempt from intellectual property and be open to all? Jennifer, and thanks for hosting the session, first of all. So, I believe there's been a certain mystique that has developed around the concepts of machine learning, of deep learning, of artificial intelligence that really has changed the perception of what it really is. And I can tell you that the more Mayo Clinic uses artificial intelligence, machine learning, and we use it for managing inventory, managing COVID supplies, diagnosing and helping treat patients, the more I realize how hard it is to define it and to define what it is. More and more, I think we should be thinking of AI as one more data analytic tool, albeit, of course, a very powerful one. And really, in healthcare, what results from training machine learning model tends to be either a new diagnostic tool for physicians or a new piece of information to aid in the diagnosis. So really, it ends up being not unlike perhaps another tool like an MRI machine. And of course, we know that these algorithms need to get approval in the US. They go to the FDA because it's a software as a device. And therefore, unless we think differently about whether all devices and medicines should be exempt from my IP protection, I do believe that it would be very hard, probably impossible and inconsistent to differentiate between one and the other. I think it's up to inventors to do so. Having said that at Mayo Clinic, we've chosen to donate some of our AI algorithms to the public. A quick example, we're working on an algorithm to diagnose COVID-19 from an electrocardiogram. We know we can perhaps even do it and determine severity. It's scalable. You could put it on a cell phone, so we want to make it available. But overall, I think it really has to be up to the people developing it to determine whether it should be IP protected or open source. Okay, thank you. So the Mayo Clinic is working with Google as your partner on your approach to collecting health data. Google is of course one of the most important actors in the AI field, but its role in collecting health care data has come under fire. Today, during this session, the forum is announcing an action plan to develop a multilateral approach to responsible data-driven AI that would include government, business, academics, and civil society, essentially bringing everyone around the table to figure out how to ensure that there is consumer trust and confidence in AI. Do you think this kind of system would increase confidence in data privacy and serve as a good alternative to heavy-handed government regulation? That's a really great question. So let me start off by saying our partnership with Google, which has ended up being a very strong partnership, is not about data collection. Our partnership with Google is two-fold. The first is we want to make sure we can protect Mayo Clinic's data. We've been given our data by our patients over 150 years. We're very aware that you can lose that trust in seconds, and therefore we want to make sure we can protect it and put it in a safe place. The second, the more exciting part is combining our clinical expertise with their cloud computing, their analytics capability, so we can accelerate health care innovation. And so what we've done is put things into our platform where we only have access to the platform and then invite others in. We can invite Google in, we can invite inference in, we can invite medically home, many other partners, and every time they only can see the data, they can't export it, so we're remaining in control on it. But of course, we derive benefit for our patients. We've promised our patients we'll innovate, and therefore there's a risk of not doing something together with the risk of doing something. Now, you're right. We do need to put some safeguards because if we don't, other people will, and so we have to continue to be very careful about this. Of course, I haven't seen the details of the proposal. It will have some merit without a doubt because we do want to be able to move forward, but understand also that different countries, different regions have different rules. So it's more, I think, about putting the standards that then can be shifted as we learn more, rather than creating yes, no kind of decisions that will actually hurt the innovation in the field. Overall, I can tell you that this is a field that is massively transforming health care in a good way. Radiation oncology is something that a lot of people don't have access to, and that's often because it takes hours and hours and hours to create a treatment plan. We've been able to use AI to massively reduce that time, create a more accurate treatment plan, and now make it available. And it's that kind of discovery that whatever standards we put in place, we have to make sure that they can continue to happen at the same time, remaining true to what is so important. People trust us with their data. We have to respect it. Thank you very much, Dr. Faruja. Let's turn now to Mr. Correa, the executive director of the South Center. How do you think we can achieve global consensus on who owns an individual's health data? Could you please walk us through how different legal systems are approaching this topic? And in your view, is a global multilateral approach necessary? Thank you, Jennifer, for hosting this meeting and for the questions. Certainly there is today a very major fragmentation about how health data and other data are legally treated. I would say that this data is kind of a limbo. For instance, in one decision by the Supreme Court of Canada, it was held that health data do belong to the doctors and not patients. Though the patient may have a control over the data, the concept was that they are owned by the doctors or the clinics and not the patients themselves. This is not the same in other jurisdictions. So there is a major effort that the international community should do. We should show that the collection and processing of data takes into account these different legal views and there is a general understanding how to deal with this issue. Otherwise, all these powerful tools will not be able to collect the necessary inputs and provide the benefit, too. In this respect, I think it's important to know that there is a major affinity today in the capacity among the doctors and developing countries in terms of collecting and processing data. And as you mentioned in your introductory remarks, it's very important that any system, any regime that is developed in the future, does take into account the quality in access to the benefits of the systems that are developed. So this is a very complex question. I cannot go through all different legal systems at this point, but we really need to look at these carefully. It's a very, very complicated situation, but there is a need to look at international agreement, international understanding of how to deal with data ownership or other type of rights. So let's move, that's a perfect segue into the next question. Let's look at the question of giving some form of value back to individuals in exchange for the data. So the forum's data for Common Purpose Initiative is a new way of looking at data governance that would improve AI by feeding it more and better data, but also benefit individuals. The idea is that devices collect information about individuals' health and lifestyles and then automatically encrypt, anonymize, and transmit along with digital rights management rules. The user would set the permission for what the data can be used for and it could be used for the public good or used for a specific commercial purpose. And if that is the case, the user could be paid. Do you think that this kind of system can work and are you in favor of it? Jennifer, this requires more exploration analysis. It's an interesting idea and I look forward for the outcomes of this project. There are very complex questions to be addressed. One of them is, as I mentioned before, who actually owns the data? Is there an ownership? Data I created, the basis of transaction is the source of the data and a supply of some technology. For instance, when you buy something with your credit card, you are not creating yourself as a user the data, but there is another entity which is participating. So who has the right to control the data? Who owns that? A difficult question. For instance, the case of sensors, which are used in many areas, do the data belong to the manufacturer of the equipment or do they belong to the user of the equipment? So all these needs to be clarified and there may be different approaches. One is this concept of ownership. The other one that you have already mentioned is to consider data as a public good, as a global public good. In fact, there has been an important report produced by a UN high-level panel which was set up by the UN Secretary-General about the digital cooperation. This panel has suggested the concept of digital global public goods in relation to data. But for other countries, data should be subject to sovereign rights because in the end, data have a value like any other asset. And therefore, it is also a very important objective for government to ensure that data are collected and processed under the sovereignty of the territory where these processes take place. So all these approaches need to be taken into account in order to find on the basis of dialogue and the international community involving on taking into account the use of developing countries to find the solution that will allow the use of these powerful tools in a manner, as you mentioned, that is fair and ensure access to all the benefits to all, in particular, developing countries and poor people, vulnerable populations. Are you familiar with different approaches that try to ensure that AI-derived data represents and is designed to help and be affordable to members of all races and ethnic groups and ensure that low-income countries or communities don't get left behind? Are there any approaches that are being tried that could help resolve this imbalance? Well, maybe one of the approaches that I mentioned is suggested by the panel on digital cooperation about the concept of global public goods. But one of the problems that was already suggested is about the appropriation of artificial intelligence systems. The number of patents in relation to artificial intelligence has increased dramatically in the last decade. There was, in fact, a study made by the UK office that has been an increase of around 400% in the last decade in terms of patent applications for artificial intelligence. So, if artificial intelligence tools are subject to appropriation to legal monopolies which are created by patents, then the data which are derived from these systems will also be subject to the control of the owners of the systems. And this is something we need to address in a very straightforward manner. There will be no access to all monopolization is allowed as already being suggested. Therefore, this need to look at this question more carefully and ensure that these technologies are available more widely. In the past, for instance, when the aircraft industry was born in the United States, there was a mandatory patent pool created in order to ensure that all manufacturers or users may have access to these technologies. And this is perhaps one model. There may be other models, but there is need to ensure that these technologies are not monopolized because this will lead in the end to the monopolization of the data which are derived from the systems. Yes, absolutely. Minister, would you like to add anything to that? Well, just to add to the points that have been made, of course, as I said, we all recognize that people must have access to their privacy, but now we're looking at something beyond individuals, although we've got to respect individuals' rights. And this therefore means that most countries must be able to develop their own ways of protecting people's privacy, just like the EU has done. Here in South Africa, we've passed our protection of the Protection of Personal Information Act, the POPIA Act that we have produced. And it specifies on what can be done with somebody's rights, personal information, and what cannot be done. But as I said, as we introduce all these changes because we are facing different challenges every day that therefore requires an individual to seize just looking at herself or her freedom, but looking at the livelihoods of the country and the economy thereof, which is why it's important that you mobilize support. Every decision that you want to make as government, you make sure that you bring people on board to say, now we require to develop these innovations or these interventions, for example, for a health-related matter. And therefore, if we are going to introduce AI boards, how do you make sure that at least bring in the human personnel to make sure that they can at least feel the look, you know, because even if we have them, they still can't do certain things, which is why we always emphasize as much as the world is saying, let's look into developing and ensuring that we encourage our young people to focus on STEM subjects. But we also add that creative side that don't forget the artistic side, because it's important that they also appreciate and understand the culture of that area. For a robot to be able to respond to my mother who is in a village, it's different to respond to me, who is in a suburb area, where things at least are advanced. Therefore, the importance of promoting languages so that they can interact freely, we're not only looking at a particular technology that we must then just extract data from, we're looking into addressing a challenge of the holistic person. That's it, Jennifer, thank you. Okay, thank you so much, Minister. And I want to thank all of the panelists for a very stimulating discussion.