 Okay, I'm very grateful for the opportunity to speak on this occasion In my talk I will be addressing a theme that has repeatedly come up during the conference namely the impact of big data on the social sphere I will not be talking so much about cities, but rather about society How this all translates to cities should be clear from what I say I Will not address the many opportunities that modern information technologies Bring to the social world, but rather also be critical. Unfortunately What I consider the most fundamental challenge and this challenge results from the rise of a network society That is from a society in which increasingly everything is linked with everything else Almost certainly in the near future the internet will expand from a network Linking just computers and their users to a network Connecting ever more areas of the social and the physical world. In fact, this is already happening Cell phones cars Buildings devices in your home at work everything is equipped with sensors Interfaces radio transmitters and will thus be integrated into the larger internet So this internet of people and things The integration of the physical and the social world into one single network permits on an unprecedented scale The collection of ever larger data sets correlating all parameters of human life. I Will argue that this allows maybe for the first time in history to establish a Substantial body of causal knowledge in the social sciences The internet and its extension to the internet of things renders obsolete Most of the traditional arguments against causal knowledge in the social sciences Namely the algorithms are much better equipped than human brains to handle the complexity of social phenomena and The internet and the internet of things increasingly allow for systematic and controlled experimentation in the social room Now owing to the very nature of causality this body of knowledge Which is being amassed by machines in the machines will allow for an increasing Predictability and control over the social sphere So let me start my argument with an analogy the analogy between Society and another complex and chaotic system the weather So what you see in the graph on the left is a depiction of the considerable improvement of weather forecasts over the last decades the different lines denote forecasts for three five seven and Ten days both for the northern hemisphere the upper lines in the southern hemisphere the lower lines respectively Essentially the slope denotes the improvement in weather predictions So what you see is a slow and gradual change which nevertheless means that for example today five day Predictions are as good as three-day predictions were ten years ago So essentially there are three different factors. You see them on the right That have led to this increase in accuracy more data increase computing power and better models But since the change is happening so slowly for many of us this quite spectacular improvement has gone largely unnoticed Now remarkably we are facing a similar situation today with respect to social phenomena Over the past decades enormous data sets have been collected mostly by large internet companies And this data is analyzed by ever more powerful computers employing automated modeling techniques Quite plausibly then these data sets allow for an improved ability to make predictions about social phenomena predictions which are getting better just as imperceptibly as weather predictions are Google for example is engaged in such social forecasting Trying to predict the movement and needs of internet users on this Google's trends platform The search history of specific terms can be looked up like in this case Internet of things and how the term Correlates with geography to see on the right or with news items the letters you see on the left Then Google's algorithms also hypothesize about the popularity of search entries in the future The dotted extension of the line Aiming at predictions in the as yet limited social world of the internet Now here's a simple formula that tells you how to turn the ability to predict into an ability to control And this formula has been known Speaking in London speaking in England This formula has been known at least since the times of Francis Bacon the intellectual father of the royal society and long Considered the founding figure of modern scientific method Bacon famously wrote that human knowledge and human power meet in one and he explicitly Refers to causal knowledge For where the cause is not known the effect cannot be produced Thus if the causes of an unknown If the causes of a phenomenon are known it becomes predictable if in addition we can manipulate these causes Then we also have control over the phenomenon plausibly the internet and the internet of things allow for just that for an increasing Ability to intervene from the distance on a large scale on the relevant causal parameters And then this new kind of control if used Systematically can drastically reduce our freedom to choose and thereby have the stupefying effects that Richard Senate has been talking about So here's an example that such mechanisms have successfully been put to work That large correlated data sets indeed allow for better prediction and control in the social sphere James Sorovicki a columnist of the New Yorker magazine Recently described how American political campaigns for example the 2008 Obama campaign employ a custom built constantly evolving algorithm that incorporates hundreds of variables in order to predict any given voters allegiance and level of enthusiasm and He continues that to campaign operatives It sometimes seemed that the algorithm actually knew what voters thought before the voters themselves did So this is a long and complex story, of course But there seems to be good reason to believe that on the basis of large data sets It was possible to identify some costly relevant parameters which subsequently could be manipulated Apparently this allowed for a small But definite degree of control to make people vote or even influence for whom so if you're interested in the details Just take a look at this recent book by Sasha Isenberg the victory lap, which you see the cover on the left So here's an obvious Objection how can we be controlled by these alleged laws given human freedom given that we are ultimately free in our choices and this apparent contradiction was already noted in the 19th century when for the first time Considerable amounts of social data were collected if still by hand and not yet by machines and by much of our present day Statistical tools were developed at the same time There was a virulent debate to what extent human actions are determined by social boundary conditions If the social sphere allows for cause loss just as the natural sciences do Is it true for example as out of cataly one of the major figures in this debate? Claim that it is society that prepares the crime and a guilty man's crime is just the fruit of the circumstances in which he finds himself In the 19th century the debate on how much human behavior is determined by cause loss ended without result mainly because all Candidates for such laws turned out much less stable and reliable than originally expected There just was not enough data, but maybe there is now To what extent can the enormous data sets of the internet be used for control of the social sphere? It is not so much an issue if we are free or not but rather at what instances we are free to choose and to which extent we are free Free that is determined not by external conditions, but by our own internal conscious deliberations So let me thus conclude As imperceptibly as weather forecast half improved in the recent decades a new kind of social science has arisen that Apparently can only exist in these large data centers as the one of Google which you see on right It may be called a social science in silico in the computers And it is made possible by the large data sets of the network society which increasingly linked every aspect of human life This novel science this novel kind of science Comprises an almost infinite number of dynamically evolving correlations and cause laws as other sciences It allows for predictions for explanations and for a certain degree of control But in other respects the science is unlike any other science that is currently taught in schools or at universities It is the science that has to stay in its digital cage as it does not fit into the limited and Static environment of conventional textbooks or even our human brains So already today, but increasingly so in the future The science will allow for the control of certain parts of the social world And it is no exaggeration to state that slowly a new instrument of power has arisen and Keeps rising and this situation Requires a debate on how this power should be used and who should be allowed to use in So to make it clear this is not about privacy It is not about personalized data rather it is often irrelevant if the data is linked to a specific name or not So we need this debate on how this social science in silico can be employed for the human good the status quo Seems at least problematic that access so far is restricted to a small number of high-tech companies Who largely use it for their economic success? Thank you You