Nell is an engineer, entrepreneur, and futurist thinker who grew up in Northern Ireland. She has a longstanding interest in the philosophy of technology, and how extensions of human capacity drive emerging social trends. Nell lectures globally on Machine Intelligence, AI philosophy, Human-Machine relations, and the Future of Human Society, serving as Associate Faculty of AI Robotics at Singularity University.
It was a great pleasure to discuss with her in this amazing interview about current situation of COVID-19 and the technology that can be useful to prevent situation like this from happening again.
So, Nell, considering Artificial Intelligence offer us powerful new ways to solve different problems, is there any chance we could use new technologies such this for beating COVID-19
Sure, there are many opportunities to prevent this kind of situation from happening again. So, AI can help us to make sense of every complex and chaotic situation and to find patterns hidden in all kinds of data.
Essentially our universe is full of correlations – people with big hands tend to also have big feet; One side of somebody’s face tends to look like a reverse image of the other side of somebody’s face. Those are simple and obvious correlations, but our universe has many very hidden or subtle correlations which we don’t immediately notice. We might have an intuition about something we might have gut instinct, but we may not be able to explain why we have that hunch, that impression.
But AI can help us to locate this kind of hidden correlations and things and that is almost like superpower. In some ways it’s the closest thing to magic in the world today. Because, an individual point of data is not terribly interesting, but if you bring information together from different sources a pattern come emerge. It’s a bit like those children’s puzzles where as you connect the dots an image begin to appear. And so, just as a picture helps to give text a context around it, or it helps to explain things in different way, when we put lots of different types of information together, AI can help us to cross-correlate, to find hidden connections between things.
A lot of the information that we are putting out into the world might be people sitting on the toilet and writing tweets, but even that information can give you some insights. We have such an abundance of information that we can put different elements of data together and uncover all kinds of new information that we would never otherwise discover.
For example, if lots of people happen to eat at the same fast-food stand, and we know from geolocation that they were all in that area, and then we notice those people don’t leave the house for couple of days, then we can start to guess that for example that food from fast-food stand is making people sick. Even if nobody reports, but they suspect that that fast food stand is selling bad food or even if the people themselves may not realize what is that exactly that made them sick. You can uncover this kind of information using machine intelligence and big data.
We could apply similar methods for figuring out who is likely to be at risk of contracting the virus or might have been in contact with other people etc. That can provide us with early warning systems. For example, there is a digital thermometer manufacturer (Kinsa) and they collect data they collect people’s temperature readings and they create a live map out of that. From this data you can build a picture in real-time of where people are having fevers, and the degree of fevers as well and you can correlate that real-time information with historical information to know that whether its likely to be a new disease or whether it’s just seasonal flu. And for example, this information is telling us that Florida in the US has a big problem with fever, there are far more fevers in Florida than usual in this time of the year, and that’s probably correlated with events like spring break where recently lots of people went out in Florida to party because they didn’t close beaches, and now Florida has a big problem with COVID-19.
So that’s another example of blending of the Internet of Things, these digital thermometers. But lots of people aren’t even aware that these things are collecting information and sending them off to the cloud. They think it’s just going to their phones and they don’t realize that that actually is going to be aggregated and anonymized as well. So, these kinds of tools can give us tremendously powerful insights into pandemics, both population level as well as on the personal level.
Many people have smartwatches that are capable of tracking your heart rate, maybe even pulse oximetry (how much oxygen you have in your bloodstream), and I’m actually helping another group in Belgium who are working to analyze heartrate data to uncover what’s called a fbril signature. It is basically a very slight change in somebody’s Electrocardiogram (ECG) from their heart which indicates that somebody could be in what’s called prodromal phase the point where you are infected but you don’t yet feel bad. In this phase, there are still small changes in the body such as the ECG of the heart and so it is possible to use machine intelligence to analyze that. Thus, if you have that heartbeat information in real-time from smartwatch, you can get an alert, in theory, which says “Watch out, looks like you are about to develop fever, so, therefore, please isolate yourself and make sure you have plenty of medicines and things if you think you need those because we have calculated that in 36 hours you are going to feel very bad”.
At the moment these kinds of smartwatches are still relatively expensive but this kind of heart rate monitor, you sometimes wear them on your chest and they are not very expensive actually, and the prices of these things are coming down all the time. If it comes down to it, we might see, maybe not in this pandemic but another one, we might see government or big tech companies for example release this kind of application so the people know even if they don’t have symptoms, they know they are likely to have this signature in their health. That would be amazing, a distributed early warning system for those about to develop illness. We could transform the infectious illness and pattern them in our society in just a few years by slashing transmission rates (R0).
We have seen in previous months that there are a lot of new services or even products for example there is one COVID -19 website where you can discover cases and when you click on some specific country then you can see the number of people who are active with COVID-19 , a number of people who died and a number of people recovered. Also in Singapore, they developed an application that shows, for example, if we went on coffee two weeks earlier and now I realize that I have COVID-19 , then app remembers all people with whom I were in contact and inform them that I’m active and that means those people know they are at risk and that is great.
How can machines understand people’s needs and how can we ensure that human side isn’t lost in all these algorithms, especially if we use AI in covid-19 diagnostic purposes
I find the Singapore response to be absolutely wonderful, very commendable they have tests that have very fast turnaround, I think the fastest test they can do is now 7 minutes which is amazing and that makes it practical to have tests for everyone as soon as they enter the country, for example. You can sit in the plane, and somebody will come around and test you, if you get the all-clear then you can go.
It is a little bit tricky because in recent years, there have been a lot of regulations against use or “abuse” technology, or health data (for very good reasons). Companies like Deep Mind were given access to huge amounts of health information in UK national health system and they faced a lot of questions about that because a lot of people thought that the degree of access that was given might not have been ideal, or perhaps a little bit too open but now we can have the flip side of the problem where we have an urgent need to have lots of data on people’s health and people’s location as well.
And it also means that, even if we are able to do that, maybe new regulations come out and bend the rules in this kind of situation which I think might be helpful and warranted. However, we still have to be very sensitive with how we do these things. We don’t want to accidentally enable people to be identified. For example, there was a patient in Korea and that one patient is basically responsible for the whole mess in Korea because he was a superspreader. The public do not know who that person is, however. It is very important for people to stay anonymized because if we look at the opposite situation, there was a lawyer in New York who met with a whole bunch of people to sign documents ( he probably sat with them to look at the papers and he maybe shared pen to sign things) and he infected over 50 people. But the problem is that his name is out there and it is out there forever. It is not his fault that he got sick, he wasn’t even aware necessarily that he was ill because people can have COVID without any symptom. It probably is not his fault and yet his name will forever have that attached to him.
And so it is very important that if this kind of early warning systems or contact management system if those are in place, that anonymity is preserved. That is something that must remain extremely strong and it might require a little bit of human oversight as well to ensure that nobody can be traced easily.
For example, it’s safe to say there was somebody at a cafe on a certain date and at a certain time, but if you say this café at this time on this date, people can maybe reverse engineer that information and turn it to identify a certain person. So, these things have to be done very sensitively and very carefully. However, they can have tremendous value, and there is more work to be done to create safe rules around this, and I would like to contribute to making better rules. I think if we need to create a special rule on health data to do that, then we should do so, but only for a very specific period temporarily for say 6 months.
Countries always says about policy lag because this is new situation for countries as well and they need time to make these new relations and to put it on operation.
Do you think that, is it important for IT companies and companies that develop new technologies to be aware of importance of privacy for developing such applications and such products which can really contribute to preventing and to minimise consequences of this virus, for example.
Yes, the crises of this moment demands that legislation be created quickly. It is therefore easy for things to go wrong, for accidental loopholes to be left in there in a why which might enable somebody to abuse something. If any extraordinary exceptions are brought in by new legislation, or emergency orders issued due to pandemic force majeure, it *must* have a fixed timeout within 12 months – a sunshine clause which cannot be extended, except by proposing the entire legislation once again. The pandemic must not be used as an excuse to erode privacy long-term or to create a foundation for lasting authoritarianism post-crises.
I wish that governance were more proactive, we knew about this situation publicly by at least mid January and we didn’t start reacting to it until a long time later. To be fair we haven’t had a pandemic at this kind of scale at least a hundred years. A lot of people worried about swine flu in 2009 and a governments reacted to that quite strongly, but not too much came of it. It was a case of “crying wolf”, which was unfortunate because eventually and inevitably the wolf has indeed come.
How the implementation of AI can help the business to minimize the effects of COVID 19 economic repercussions
These are very difficult and challenging times. I think that only good business to be in right now is plastics and maybe pharmaceutical as well, some forms of remote entertainment. There are definitely things we can do, that can make a big difference. For example, there are groups online and projects which are working leverage technologies such as 3D printing to make face shields and masks which is amazing because there is such dire need for Personal Protective Equipment (PPE).
3D printing has been talked about for a long time, but without a “killer app” that makes it a must-have. This killer app has been found saving our lives and protecting our health.
We can blend 3D printing with Artificial Intelligence because we can use generative design technique where basically you give your objective to the machine for example, a bicycle frame with minimal mass, material cost and maximum strength. Machine intelligence can evolve versions through “genetic” process which are progressively more sophisticated and more optimized.
Then we can 3D print that object out, because with 3D printing complexity becomes essentially free, it doesn’t matter how complex something is, it is just printed layer by layer, and it is basically at same cost to make something very complex or intricate.
This is going to make a revolution for local production. I think in many ways that the era of everything being made in China or offshore is to a significant degree is going to change. We are going to see a lot of governments, societies rehousing factories back in local places, local fabrication centres to make things – things customized for efficiency for minimal material and maximal strength. They are going to be customized to fit people’s bodies if necessary and they are going to be customized in terms of aesthetics as well. There are tremendous new opportunities in those kinds of areas. Technology has its stride and now people are waking up to it, it is crossing from a few “early adopters” into the mainstream, and that’s going to make a big difference going forward.
Nell’s speech was truly inspirational, so as a conclusion she said: It is fantastic and inspirational that so many people from all around the world are coming together. This is the first time that all humanity has been united in one fight. This is something that we all have a stake in, a very personal stake. This is a very special time in history, and it may become our finest hour as a species. These are indeed tricky times, but interesting times and, the silver lining for this cloud might be a good time to create new ways of thinking about the world, new institutions, new systems.