There are two big global trends. One is big data. That means in the next ten years we'll produce as many data, or even more data than in the past 1,000 years. The other trend is hyperconnectivity. That means we have networking our world going on at a rapid pace; we're creating an Internet of things. So everyone is talking to everyone else, and everything becomes interdependent. What are the implications of that? Well, first of all, of course, we have wonderful new gadgets like Facebook, for example, so we can network with each other. We have new services, new opportunities, and that is just fantastic.
But on the other hand, it turns out that we are, at the same time, creating highways for disaster spreading. We see many extreme events, we see problems such as the flash crash, or also the financial crisis. That is related to the fact that we have interconnected everything. In some sense, we have created unstable systems. We can show that many of the global trends that we are seeing at the moment, like increasing connectivity, increase in the speed, increase in complexity, are very good in the beginning, but (and this is kind of surprising) there is a turning point and that turning point can turn into a tipping point that makes the systems shift in an unknown way.
It requires two things to understand our systems, which is social science and complexity science; social science because computers of tomorrow are basically creating artificial social systems. Just take financial trading today, it's done by the most powerful computers. These computers are creating a view of the environment; in this case the financial world. They're making projections into the future. They're communicating with each other. They have really many features of humans. And that basically establishes an artificial society, which means also we may have all the problems that we are facing in society if we don't design these systems well. The flash crash is just one of those examples that shows that, if many of those components — the computers in this case — interact with each other, then some surprising effects can happen. And in that case, $600 billion were actually evaporating within 20 minutes.
Of course, the markets recovered, but in some sense, as many solid stocks turned into penny stocks within minutes, it also changed the ownership structure of companies within just a few minutes. That is really a completely new dimension happening when we are building on these fully automated systems, and those social systems can show a breakdown of coordination, tragedies of the commons, crime or cyber war, all these kinds of things will happen if we don't design them right.
We really need to understand those systems, not just their components. It's not good enough to have wonderful gadgets like smartphones and computers; each of them working fine in separation. Their interaction is creating a completely new world, and it is very important to recognize that it's not just a gradual change of our world; there is a sudden transition in the behavior of those systems, as the coupling strength exceeds a certain threshold.
I'd like to demonstrate that for a system that you can easily imagine: traffic flow in a circle. Now, if the density is high enough, then the following will happen: after some time, although every driver is trying hard to go at a reasonable speed, cars will be stopped by a so-called ‘phantom traffic jam.' That means smooth traffic flow will break down, no matter how hard the drivers will try to maintain speed. The question is, why is this happening? If you would ask drivers, they would say, "hey, there was a stupid driver in front of me who didn't know how to drive!" Everybody would say that. But it turns out it's a systemic instability that is creating this problem.
That means a small variation in the speed is amplified over time, and the next driver has to brake a little bit harder in order to compensate for a delayed reaction. That creates a chain reaction among drivers, which finally stops traffic flow. These kinds of cascading effects are all over the place in the network systems that we have created, like power grids, for example, or our financial markets. It's not always as harmless as in traffic jams. We're just losing time in traffic jams, so people could say, okay, it's not a very serious problem. But if you think about crowds, for example, we have this transition towards a large density of the crowd, then what will happen is a crowd disaster. That means people will die, although nobody wants to harm anybody else. Things will just go out of control. Even though there might be hundreds or thousands of policemen or security forces trying to prevent these things from happening.
This is really a surprising behavior of these kinds of strongly-networked systems. The question is, what implication does that have for other network systems that we have created, such as the financial system? There is evidence that the fact that now every bank is interconnected with every other bank has destabilized the system. That means that there is a systemic instability in place that makes it so hard to control, or even impossible to control. We see that the big players, and also regulators, have large difficulties to get control of these systems.
That tells us something that we need to change our perspective regarding these systems. Those complex systems are not characterized anymore by the properties of their components. But they're characterized by what is the outcome of the interactions between those components. As a result of those interactions, self-organization is going on in these systems. New emergent properties come up. They can be very surprising, actually, and that means we cannot understand those systems anymore, based on what we see, which is the components.
We need to have new instruments and tools to understand these kinds of systems. Our intuition will not work here. And that is what we want to create: we want to come up with a new information platform for everybody that is bringing together big data with exa-scale computing, with people, and with crowd sourcing, basically connecting the intelligence of the brains of the world.
One component that is going to measure the state of the world is called the Planetary Nervous System. That will measure not just the physical state of the world and the environmental situation, but it is also very important actually that we learn how to measure social capital, such as trust and solidarity and punctuality and these kinds of things, because this is actually very important for economic value generation, but also for social well-being.
Those properties as social capital, like trust, they result from social network interactions. We've seen that one of the biggest problems of the financial crisis was this evaporation of trust. It has burned tens of thousands of billion dollars. If we would learn how to stabilize trust, or build trust, that would be worth a lot of money, really. Today, however, we're not considering the value of social capital. It can happen that we destroyed it or that we exploit it, such as we've exploited and destroyed our environment. If we learn how much is the value of social capital, we will start to protect it. Also we'll take it into account in our insurance policies. Because today, no insurance is taking into account the value of social capital. It's the material damage that we take into account, but not the social capital. That means, in some sense, we're underinsured. We're taking bigger risks than we should.
This is something that we want to learn, how to quantify the fundaments of society, to quantify the social footprint. It means to quantify the implications of our decisions and actions.
The second component, the Living Earth Simulator will be very important here, because that will look at what-if scenarios. It will take those big data generated by the Planetary Nervous System and allow us to look at different scenarios, to explore the various options that we have, and the potential side effects or cascading effects, and unexpected behaviors, because those interdependencies make our global systems really hard to understand. In many cases, we just overlook what would happen if we fix a problem over here: It might have unwanted side effects; in many cases, that is happening in other parts of our world.
We are using supercomputers today in all areas of our development. Like if we are developing a car, a plane or medical tracks or so, supercomputers are being used, also in the financial world. But we don't have a kind of political or a business flight simulator that helps us to explore different opportunities. I think this is what we can create as our understanding of society progresses. We now have much better ideas of how social coordination comes about, what are the preconditions for cooperation. What are conditions that create conflict, or crime, or war, or epidemicspreading, in the good and the bad sense?
We're using, of course, viral marketing today in order to increase the success of our products. But at the same time, also we are suffering from a quick spreading of emerging diseases, or of computer viruses, and Trojan horses, and so on.We need to understand these kinds of phenomena, and with the data and the computer power that is coming up, it becomes within reach to actually get a much better picture of these things.
The third component will be the Global Participatory Platform. That basically makes those other tools available for everybody: for business leaders, for political decision-makers, and for citizens. We want to create an open data and modeling platform that creates a new information ecosystem that allows you to create new businesses, to come up with large-scale cooperation much more easily, and to lower the barriers for social, political and economic participation.
So these are the three big elements. We'llfurthermore build exploratories of society, of the economy and environment and technology, in order to be able to anticipate possible crises, but also to see opportunities that are coming up. Those exploratories will bring these three elements together. That means the measurement component, the computer simulation component, and the participation, the interactiveness.
In some sense, we're going to create virtual worlds that may look like our real world, copies of our world that allow us to explore polices in advance or certain kinds of planning in advance. Just to make it a little bit more concrete, we could, for example, check out a new airport or a new city quarter before it's being built. Today we have these architectural plans, and competitions, and then the most beautiful design will have win. But then, in practice, it can happen that it doesn't work so well. People have to stand in line in queues, or are obstructing each other. Many things may not work out as the architect imagined that.
What if we populated basically these architectural plans with real people? They could check it out, live there for some months and see how much they like it. Maybe even change the design. That means, the people that would use these facilities and would live in these new quarters of the city could actually participate in the design of the city. In the same sense, you can scale that up. Just imagine Google Earth or Google Street View filled with people, and have something like a serious kind of Second Life. Then we could have not just one history; we can check out many possible futures by actually trying out different financial architectures, or different decision rules, or different intellectual property rights and see what happens.
We could have even different virtual planets, with different laws and different cultures and different kinds of societies. And you could choose the planet that you like most. So in some sense, now a new age is opening up with almost unlimited resources. We're, of course, still living in a material world, in which we have a lot of restrictions, because resources are limited. They're scarce and there's a lot of competition for these scarce resources. But information can be multiplied as much as you like. Of course, there is some cost, and also some energy needed for that, but it's relatively low cost, actually. So we can create really almost infinite new possibilities for creativity, for productivity, for interaction. And it is extremely interesting that we have a completely new world coming up here, absolutely new opportunities that need to be checked out.
But now the question is: how will it all work? Or how would you make it work? Because the information systems that we have created are even more complex than our financial system. We know the financial system is extremely difficult to regulate and to control. How would you want to control an information system of this complexity? I think that cannot be done top-down. We are seeing now a trend that complex systems are run in a more and more decentralized way. We're learning somehow to use self-organization principles in order to run these kinds of systems. We have seen that in the Internet, we are seeing t for smart grids, but also for traffic control.
I have been working myself on these new ways of self-control. It's very interesting. Classically one has tried to optimize traffic flow. It's so demanding that even our fastest supercomputers can't do that in a strict sense, in real time. That means one needs to make simplifications. But in principle, what one is trying to do is to impose an optimal traffic light control top-down on the city. The supercomputer believes to know what is best for all the cars, and that is imposed on the system.
We have developed a different approach where we said: given that there is a large degree of variability in the system, the most important aspect is to have a flexible adaptation to the actual traffic conditions. We came up with a system where traffic flows control the traffic lights. It turns out this makes much better use of scarce resources, such as space and time. It works better for cars, it works better for public transport and for pedestrians and bikers, and it's good for the environment as well.
There's a new kind of socio-inspired technology coming up, now. Society has many wonderful self-organization mechanisms that we can learn from, such as trust, reputation, culture. If we can learn how to implement that in our technological system, that is worth a lot of money; billions of dollars, actually. We think this is the next step after bio-inspired technology.
The next big step is to focus on society. We've had an age of physics; we're now in an age of biology. I think we are entering the age of social innovation as we learn to make sense of this even bigger complexity of society. It's like a new continent to discover. It's really fascinating what now becomes understandable with the availability of Big Data about human activity patterns, and it will open a door to a new future.
What will be very important in order to make sense of the complexity of our information society is to overcome the disciplinary silos of science; to think out of the box. Classically we had social sciences, we had economics, we had physics and biology and ecology, and computer science and so on. Now, our project is trying to bring those different fields together, because we're deeply convinced that without this integration of different scientific perspectives, we cannot anymore make sense of these hyper-connected systems that we have created.
For example, computer science requires complexity science and social science to understand those systems that have been created and will be created. Why is this? Because the dense networking and to the complex interaction between the components creates self-organization, and emergent phenomena in those systems. The flash crash is just one example that shows that unexpected things can happen. We know that from many systems.
Complexity theory is very important here, but also social science. And why is that? Because the components of these information communication systems are becoming more and more human-like. They're communicating with each other. They're making a picture of the outside world. They're projecting expectations into the future, and they are taking autonomous decisions. That means if those computers interact with each other, it's creating an artificial social system in some sense.
In the same way, social science will need complexity science and computer science. Social science needs the data that computer science and information communication technology can provide. Now, and even more in the future, those data traces about human activities allow us eventually to detect patterns and kind of laws of human behavior. It will be only possible through the collaboration with computer science to get those data, and to make sense of what is happening actually in society. I don't need to mention that obviously there are complex dynamics going on in society; that means complexity science is needed for social science as well.
In the same sense, we could say complexity science needs social science and computer science to become practical. To go a step beyond talking about butterfly effects and chaos and turbulence. And to make sure that the thinking of complexity science will pervade our thinking in the natural engineering and social sciences and allow us to understand the real problems of our world. That is kind of the essence: that we need to bring these different scientific fields together. We have actually succeeded to build up these integrated communities in many countries all over the world, ready to go, as soon as money becomes available for that.
Big Data is not a solution per se. Even the most powerful machine learning algorithm will not be sufficient to make sense of our world, to understand the principles according to which our world is working. This is important to recognize. The great challenge is to marry data with theories, with models. Only then will we be able to make sense of the useful bits of data. It's like finding a needle in the haystack. The more data you have, the more difficult it may be to find this needle, actually, to a certain extent. And there is this danger of over-fitting, of being distracted from important details. We are certainly already in an age where we're flooded with information, and our attention span can actually not process all that information. That means there is a danger that this undermines our wisdom, if our attention is attracted by the wrong details of information. So we are confronted with the problem of finding the right institutions and tools and instruments for decision-making.
The Living Earth Simulator will basically take the data that is gathered by the Internet, search requests, and created by sensor networks, and feed it into big computer simulations that are based on models of social and economic and technological behavior. In this way, we'll be able to look at what-if scenarios. We hope to get a better understanding, for example, of financial systems and some answers to controversial questions such as how much leverage effect is good? Under what conditions is 'naked short-selling' beneficial? When does it destabilize markets? To what extent is high frequency trading good, or it can it also have side effects? All these kinds of questions, which are difficult to answer. Or how to deal best with the situation in Europe, where we have trouble, obviously, in Greece, but also kind of contagious effects on other countries and on the rest of the financial system. It would be very good to have the models and the data that allow us actually to simulate these kinds of scenarios and to take better-informed decisions.
Many people are asking themselves, who will benefit from the information and communication platform that we want to create? The answer is: we want to open it up for everybody, such as everybody is using Wikipedia, and everybody can contribute actually to Wikipedia. We don't want to create a machine that is giving the answers to what people should do. We want to create instruments that allow people to look at complex problems from multiple perspectives, such as we have created telescopes to explore our environment and our universe from different perspectives. So it [the information and communication platform] won't give you the answer, but it will give you more information, and then everybody will be able to be better-informed and make decisions based on personal values and priorities.
Of course, no computer can decide about those values and priorities. This is, in the end, a personal or a political or a business strategy issue. It's very important that we keep it neutral. The idea is to have an open platform to create a data and model commons that everybody can contribute to, so people could upload data and models, and others could use that. People would also judge the quality of the data and models and rate them according to their criteria. And we also point out the criteria according to which they're doing the rating. But in principle, everybody can contribute and everybody can use it.
People will get different answers, depending on their respective values and priorities. On the other hand, they will have a much better basis of discussion when it comes to the need of decision-making.
I started off as a physicist, very much driven by the curiosity to understand the universe and elementary particles and all this. But eventually I got more interested in real-life problems. In Göttingen, where I studied, there was a university where many of the quantum physicists were working, like Max Planck, for example. I got interested in social phenomena such as pedestrian crowds, and eventually also crowd disasters, which is really interesting. Why are crowd disasters happening? Nobody wants to harm anybody else. So why do things go wrong? Then I was extending this kind of work towards traffic modeling, to understand phantom traffic jams and how to control or organize these complex systems. Everybody is annoyed by traffic jams, and the question is, why are they happening and what can we do against them?
There are quite a few interesting answers, and the new approach allows one to get rid of traffic jams to a large extent. Eventually when I was at TU Dresden, which is a university in Eastern Germany, there was a huge flooding, and we extended our emergency-related studies to flooding and other kinds of natural disasters, and to disaster response management and so on. I got more and more interested in systemic risks. When I finally got to ETH Zurich, which is continental Europe's leading university, I became a professor of sociology, bringing me back to this interest in crowds and to collective behavior and opinion information, cooperation, conflict and so on.
Then basically the financial crisis, when it was coming up, was triggering my interest in bigger and global systems, because it became apparent that we didn't know enough about these kinds of systems. Nobody knew enough about them, I would say. I had the impression we knew more about our universe and elementary particles, actually, than about our society and economy. That is something I wanted to change. I think it was the right time to do so, because now there is all the data becoming available, and the computer power. We have much better theories, also, that allows us to make sense of those data. We're entering into an age where we can understand society and the economy much better, namely as complex self-organizing systems.
It will be important to guide us into the future because we are creating very powerful systems. Information society will transform our society fundamentally and we shouldn't just let it happen. We want to understand how that will change our society, and what are the different pathes that our society may take, and decide for the one that we want it to take. For that, we need to have a much better understanding.
Now a lot of social activity data are becoming available through Facebook and Twitter and Google search requests and so on. This is, of course, a huge opportunity for business. Businesses are talking about the new oil, personal data as new asset class. There's something like a gold rush going on. That also, of course, has huge opportunities for science, eventually we can make sense of complex systems such as our society. There are different perspectives on this. They range from some people who think that information communication technologies eventually will create a God's-eye view: systems that make sense of all human activities, and the interactions of people, while others are afraid of a Big Brother emerging.
The question is how to handle that situation. Some people say we don't need privacy in society. Society is undergoing a transformation, and privacy is not anymore needed. I don't actually share this point of view, as a social scientist, because public and private are two sides of the same coin, so they cannot exist without the other side. It is very important, for a society to work, to have social diversity. Today, we know to appreciate biodiversity, and in the same way we need to think about social diversity, because it's a motor of innovation. It's also an important factor for societal resilience. The question now is how all those data that we are creating, and also recommender system and personalized services are going to impact people's decision-making behavior, and society overall.
This is what we need to look at now. How is people's behavior changing through these kinds of data? How are people changing their behavior when they feel they're being observed? Europe is quite sensitive about privacy. The project we are working on is actually trying to find a balance between the interest of companies and Big Data of governments and individuals. Basically we want to develop technologies that allow us to find this balance, to make sure that all the three perspectives actually are taken into account. That you can make big business, but also at the same time, the individual's privacy is respected. That individuals have more control over their own data, know what is happening with them, have influence on what is happening with them.
That requires new kinds of data mining technologies, and it is a big challenge to open these data up for everybody, and to avoid misuse with this kind of data. How to promote responsible use of such a system, as a data commons is a public good, and public goods are often threatened by tragedies of the commons? There could be data pollution, there could be cybercrime, all these kind of things, so it's really a fundamental question. What do you have to do in order to be able to open it up for a large number of people and to have a responsible use? There are a number of underlying research questions that are really very fundamental, which we have formulated and are preparing to address.
This is also where we distinguish ourselves from big companies that are collecting huge amounts of data. Of course these companies are profit-driven and ought to be profit-driven. But there are also societal interests, and interests of private individuals that would not necessarily be taken care of by those profit-driven activities. We are talking about creating public goods, like we create public infrastructures such as schools, universities, theaters. Our language is also something like a public good. In some sense, we want to create a new data and model commons, a new kind of language, a new public good that allows people to do new things.
Just imagine every company would own ten words. We couldn't speak as we do today, using words freely. We would be very much hampered in our way of interaction. It's very important that we do have such kinds of joint platforms that allow us to interact seamlessly with each other, and based on these platforms, you can, of course, build a lot of businesses.
My feeling is that actually business will be made on top of this sea of data that's being created. At the moment data is kind of the valuable resource, right? But in the future, it will probably be a cheap resource, or even a free resource to a certain extent, if we learn how to deal with openness of data. The expensive thing will be what we do with the data. That means the algorithms, the models, and theories that allow us to make sense of the data.