Big data is easy to learn

back April 19, 2016

What is big data
Explained with 10 questions and answers

Ten questions for Viktor Mayer-Schönberger

In an interview with Vera Linß and Camilla Graubner, Viktor Mayer-Schönberger answers ten questions about big data for the magazine "tv diskurs".

We leave our digital fingerprints at companies all over the Internet. In return, we should be able to expect them to responsibly protect our data. License: cc by-nd / 2.0 / de (Flickr - CPOA)

Viktor Mayer-Schönberger, Professor at the Internet Institute at Oxford University and an expert at the Bonn Talks 2016 in conversation with Vera Linß and Camilla Graubner from the magazine "tv diskurs"

Linß / Graubner: How would you define big data?

Mayer-Schönberger: Big Data is a new perspective on reality that is based on the analysis of a large number of data points.

What can be read from the amount of data - compared to the time before big data?

At its core it is about a better understanding of the world. In the age of small data, we collected and analyzed data in order to answer specific questions that we asked ourselves. With big data, we can use the patterns found in the data to inspire us to ask new questions that we didn't even know we should be asking. For example, the computer scientist Dr. Carolyn McGregor found patterns in vital signs data in premature babies that correlate with likely later infection. This makes it possible to predict a possible illness - 24 hours before the first symptoms appear. Analyzing large amounts of data can save lives, as it does with these premature babies. More generally, it allows us to make better decisions in everyday life.

To what extent is it desirable to record everything quantitatively?

That is most desirable. The alternative would be ignorance of the world we live in - and insistence on superstition, stereotypes, ideologies and sensitivities. This ignorance has already cost too many people their lives. Europe's central contribution to human development is quite rightly the Enlightenment. Seen in this way, big data is the enlightenment for the 21st century.

"Big data is about the what, not the why," you write in your book. So can big data be seen as a supplement to classic scientific work?

Yes - and as a support. We humans will continue to search for causes, but this search for causes is extremely time-consuming and costly. It therefore makes sense to use them very consciously. Recognizing the "what" with the help of big data allows us to filter out the most interesting connections for the search for the "why". And, moreover, pragmatically deriving instructions for action from the "what": for example, in the case of premature babies, giving them medication before the probable infection fully breaks out.

How is the media changing due to big data?

That is hard to say. In comparison, the development is still at the beginning. Of course, big data is already being used today to understand consumer preferences more precisely - not only to select "suitable" content from a media offering, but also to allow it to emerge in the first place. Some have criticized this because it only reinforces our preferences and no longer confronts us with the unexpected. But it's not that simple. Because for those who value the unexpected, big data analysis will find out and also offer them the unexpected. The desire to be surprised is also predictable to a certain extent.

What risks do you see with regard to the misuse of big data?

Much has been heard about the fact that long-term storage and retention of personal data always ties us to our past and does not sufficiently reflect the ability of people to learn to develop. That's right. But beyond that, I see two further and at least as problematic dangers of abuse. The first danger is that we are tempted to link the prediction of future behavioral consequences - in particular individual responsibility - to "judge" people simply by making the prediction. Ultimately, that would be the end of free will - a terrible idea! Precisely for this reason, a legal framework is required that defines which use of data is permitted in the context of big data - and which is not (or only under very specific conditions). The second danger is that we misinterpret big data analysis, which can usually only tell us "what" is happening or will happen, and also derive a "why". We humans are shaped by seeing the world as a chain of causes and effects - and very often believe that we have recognized the causes without actually being able to do so. It gives us the deceptive feeling that we have seen through the world. Therefore, every big data analysis is always exposed to the risk of being misinterpreted by us humans, in that we attach more meaning and importance to the results of the analysis than they are due. Then we go into the dependency, even the dictatorship of the data. This must also be prevented - through clear rules, but also through competence work and educating people.

How can we ensure that individuals have room to maneuver? That it is not prejudiced by data collected on the basis of algorithms?

Among other things, by deliberately stating that we want to remain ignorant of certain questions.

There have always been prejudices or assumptions about developments and about people. What is the new quality if knowledge is gained with the help of algorithms?

There is no new quality to the algorithms. The algorithms are stupid. The insights are in the data. But even these do not have a new quality. What is new is "only" our ability to collect and analyze the data much more comprehensively. And the danger is that we humans fail to interpret the analysis results. We need ethics in the handling and use of data.

The fact that algorithms have to be regulated has been discussed for years. Do you have the impression that we have made progress?

My fear is that we are indeed beginning to regulate algorithms - and thus completely misunderstand the real challenge that lies in the data and its analysis.

How do you rate society's awareness of the dangers posed by algorithms?

It is wrong, marked by false reports about the "danger" posed by algorithms. This misunderstanding worries me.

It is reprinted with the kind permission of tv diskurs

Source: Mayer-Schönberger, Viktor: “Big Data is the enlightenment for the 21st century.” Ten questions for Viktor Mayer-Schönberger; in: tv diskurs JG 75, 2016 (1), pp. 32 - 35.

Download the interview as a PDF

To the full edition tv discourse 1/2016
We leave our digital fingerprints at companies all over the Internet. In return, we should be able to expect them to responsibly protect our data.

(Flickr - CPOA)

Previous Next
We use cookies on this website. These are used to offer you service functions and for statistical and analysis purposes (web tracking). How you can object to web tracking and further information can be found in our data protection declaration.