As an expert on AI and Big Data, Bart de Witte is fighting for patient data to be shared in a spirit of solidarity and for the greater good. In this long-read interview with //next columnist Markus Sekulla, the entrepreneur shares his view on what impact AI can have on healthcare - and what needs to be done to democratize AI for a brighter future.
I grew up in an extremely technologically advanced family. My father was an entrepreneur and had computers in his company. I think it was in 1984 when IBM had its first portable computer. And that was the first time he brought one home. So, I had the opportunity to observe and play with computers for the first time, which was incredibly thrilling. That's how it began.
I was completely captivated by computers and began programming in BASIC, then Pascal, and later Assembler. We were among the first people to use CompuServe, a forerunner of the internet, in 1990. This was a very early chance for me to interact with (at the time) cutting-edge technology, and it changed my entire life. Furthermore, it encouraged my curiosity, allowing me to have an open mind, interact with individuals all over the world, and learn or, more importantly, unlearn things ...
I believe it was because I had access to computers throughout the 1980s, when computer games were at their peak. I grew raised in a very remote area near Antwerp, Belgium. I didn't have many neighbors with whom I could play. Suddenly, I had this buddy with whom I could compete, which was a computer game, and it felt good to have this artificial friend.
While my uncle was working on expert systems for Belgian banks, I wrote a high school paper about how AI can assist African physicians make better diagnoses. And it was 1989, and the irony is that I had a poor grade because no one appeared to understand what AI was. I moved on to work at SAP and IBM, where I met individuals who were using machine learning to analyze health data. We suddenly had access to processing power, techniques, and data, allowing us to perform precisely what I spoke about in my school paper, which was democratizing and scaling medical knowledge to remote locations.
Catharsis as a media approach usually works best. At least three front pages in Der Spiegel have highlighted robots or artificial intelligence taking our jobs. Others believed Frey and Osborne's 2013 Oxford research, which predicted that AI will eliminate half of all occupations. I and many other researchers disagree with their forecasts since they were based on incorrect assumptions and provided a skewed vision of how we would employ AI in society. The study was incorrect and instilled rage, fear, and excitement - ideal content for clickbait headlines to serve our attention economy. Media amplifies stories to increase their advertisement revenue, which is boosted by the amount of hits, likes and shares.
Positive narratives that lead to a more desired future are needed. What type of future do we want our children and society to have? Do we desire a future in which technology frees us? Do we want to eliminate global health inequities in which the site of my birth won’t decide anymore whether my kid lives or dies, for example, if it has a rare genetic disease? Do we want AI's benefits to benefit society first?
Open thing is sure, building barriers around life-saving knowledge derived from our data by AI won’t allow us to achieve this desirable future. To advance, we must create medical AI in the same manner that we intended the internet to be open and available to everybody.
Today's discourse is overly focused on economics, thus the topic that has kept me occupied the last few years is how we may maximize societal benefits while supporting economic opportunity.
Simplified we can say that AI in medicine works in two ways. First, it unlocks the wisdom buried in our data. It helps in the acceleration of research and innovation. In healthcare, for example, this may entail new treatments and therapies. Second, when we use it, we can scale it almost everywhere. When much of our current technology was developed, it followed a law known as Moore's law. Someone living in Nigeria today, has access to more information than President Clinton did 15 years ago. And if they put a 3D printed optical device to it and use some machine learning algorithms, they can identify HPV infections more precisely than those who utilize pap-smear and cytology for nearly nothing. These are the things that intrigue me about technology, as it demonstrates that technology can liberate us.
We all have been empowered by open and decentralized technologies. Consider what might happen if we democratized medical AI? The democratization aspect is what motivates me since it will enable us to construct a society free of health inequities and develop new layers of economic value that compete on an experiential level. This is a realistic vision, because most disparities are founded on knowledge asymmetries. Consider the vaccination discussion at COVID19. At the time we were contemplating our third booster, just 5% of Africa had access to immunizations. I was only recently that academics published open-source vaccinations for low-income countries.
AI is already a game changer. I think even BioNTech used machine learning protocols to develop the vaccine so fast. Machine learning is already an invisible part on our scientifical and R&D layer in healthcare. It is just not visible to you and me in that sense.
Medicine will evolve into applied digital biology. It implies that we will begin to comprehend how complex human systems act on a molecular level. We shall begin to grasp the language of our own biology. What exactly is the illness? Where does it originate from? What is it? What happens on a molecular level? Why do we get cancer? And how can we, on a molecular level cure this? We are still not certain of the answer since we only comprehend a small portion of human biology. AI will allow us to comprehend all genetic levels, the entire cell structure, how they communicate, and so on. The most significant discoveries in the next ten years will all be in the field of digital and synthetic biology.
We will be able to intervene extremely early because we will understand what kinds of behavioral, lifestyle, and dietary adjustments may be applied to a body. Or we will be able to act very early with breakthrough biotech therapies, and we will no longer get sick. In the end, we may be able to avert a sickness in certain ways. Some techno-optimists believe we will be able to cure or reverse the aging process.
Exclusivity will be less of a worry for future medical advancements because many of the technologies required to create and produce these items are abundant technologies. This implies they are obeying Moore's law, which outlines how computer technology has constantly been more powerful while also becoming less expensive. These technologies have the potential to make scarcity obsolete in many industries, but it all depends on the economic foundation we develop around data and AI in healthcare.
Different leadership toward a more sustainable and fair society, as well as the breaking down of hierarchical organizations to gain speed and profit from the benefits of these exponential technologies, are some of the ingredients that will lead to a different culture of plenty. It will enable health- and life-sciences enterprises to create and apply the same successful formulae as digital players: speed, data knowledge, and an unwavering focus on the patient and user. Digital societies provide new opportunity to realign value production with long-term goals.
The existing paradigm, which is built on artificial scarcity, is a model created for inequity and contradicts the WHO Constitution, which proclaimed healthcare a core human right and has become one of the new generation's ideals. Adopting solidarity ideals and open source collaboration in data and AI will help us to achieve this abundant planet.
That is certainly my intention; it is named after Hippocrates, the father of modern medicine. The modern Western sciences which began in the 16/1700s., where open sciences. The notion of open access sprang from the medical sciences, and the ethics of medical science arose from Greek culture, with Hippocrates laying the groundwork. Hippocrates' original oath was that physicians must impart their expertise for free and without financial gain. As a result, open source was included in the original order document. But there was one distinction. It was not entirely open source since it needed to be shared with colleagues but not with patients. I adopted that mindset and felt compelled to lay the groundwork with an ethical guideline, ensuring that the information we make from data is always open to everybody.
Everything in our Western society is determined by policies and incentives. With these technology' increasing capabilities, the general practitioner may become a type of super-doctor. One may even go so far as to let others to "diagnose." Why not give these powerful tools to a pharmacist or a nurse? This is a heated but vital debate. As the best diagnostic is one that is available to everyone at all times and at extremely little cost.
COVID19 has opened the door for self-testing, and resulted in a large number of decentralised test-centers. This was only feasible because Christian Drosten's Charité Team published the PCR testing procedure as a non-patented open-protocol. What if we go a step further and enable these types of models, along with open-source AI, to help our healthcare systems increase supply? I am confident that we would witness remarkable service innovation that would produce far superior experiences than the system now provides. To achieve this aim, we should avoid erecting data barriers and establishing digital knowledge kingdoms, which would result in a digital healthcare system owned by only a few tech-feudalists. We must establish a foundation based on the ideals of solidarity, openness, decentralization, and democracy.
In 2030, we will be able to apply AI to data generated by digital tools or digital chatbots used by patients. Perhaps we can test ourselves at home and connect to digital health care providers online. And, perhaps, we were able to leverage the efficiency improvements to enhance the amount of human-to-human engagement when needed. As at the end, when we are sick, next to decision making, human touch and empathy always will play a big role.
Interview: Markus Sekulla