“The word Big data has been kidnapped by marketing” Fabrice Epelboin
Frenchweb : The term “Big data” is used everywhere. What does it mean? And, on the contrary, what does it not mean?
Fabrice Epelboin: The era of Big Data represents an order of magnitude change in the quantity and variety of data that is subject to computer processing. It also represents a change in the tools used to do this treatment. The database model that prevailed until now is no longer suitable, and leaves room for other approaches such as the famous hadoop. Of course, Big Data goes with the considerable increase in storage capacities, which means that we now keep everything that could be processed by big data, in particular data and personal traces left behind. everywhere by Internet users, which advertisers are fond of.
The analysis of this Big Data is supposed to provide decision support, or even to automate it. As is often the case, we fall back into the dichotomy of a man / machine approach versus a purely algorithmic approach, which is nothing new. This duality is a great finding in IT. But with big data, we will find more and more decision-makers facing the machine; decision-makers who will not have the slightest idea of how the machine works, which is not without asking a number of questions, in particular about the insidious political role of the technician.
Finally, it should be noted that the cost of processing this data, which was once prohibitive, is in free fall. Analyzing for the purpose of extracting value, a huge amount of data is no longer reserved for companies capable of setting up heavily funded projects; it is within the reach of the first start-up to come, as long as it has data or knows how to collect it. This last point, which at first glance seems trivial or expected, has important consequences: capital does not constitute a barrier to the entry of big data. Anyone or almost can do it, even if, as specialists like to repeat it,
Big Data is like teenage sex: everyone talks about it, everyone claims to do it, but in reality …
Is this word a buzzword used wrongly and through?
Like many tech words over the past decade, it has been kidnapped by marketing (but not only), and its meaning has been greatly altered. To this day, we can already consider that this word is moribund. This term has become a catch-all intended to give a “young and dynamic” connotation to anything and everything as long as there is technology in it. It’s a great classic, the “Cloud” has known the same semantic fate, just like the “social web” before it.
But to accuse marketing of this semantic homicide, it is all the same to go a little fast in work. This kind of semantic killing happens when a multitude of pundits, journalists and neophytes flock to a new word like misery on the poor world, it goes far beyond marketing. We could just as easily blame “social networks”, which are, much more than marketing, a designated culprit much more fashionable these days. The truth is, this kind of semantic sleight of hand for more or less marketing purposes is a classic. The word “Republic”, of which few people in the end know the real definition (the management of the common good and sovereign powers), is today the object of repeated semantic assaults on the part of the political class in its together, to the point of having been squarely sequestered by the UMP which made it a registered trademark.
The reality is that any disruption is scary, that Big Data represents a major form of disruption, well beyond the technological field, and that thus grabbing a word, and “diluting” it semantically, in a way, is also a way of familiarizing yourself and removing the anxiety-inducing aspect.
Big Data is nonetheless the emerging part of a deep disruption, with consequences to come on our political governance systems, our health system, or the social in general: all this will be profoundly modified from the simple fact of the appearance of big data.
Which sectors or companies are making the best use of it, and which are just marketing it?
Marketing can be the most effective end to big data, Google knows that. But it is indeed outside of marketing that we will find the most disruptive uses – from there to say that these uses are good, we will have to wait a bit and take a step back. What is certain is that these uses are not supervised, and that they are, for many, in the laboratories of the intelligence services or those of their subcontractors.
Mass surveillance, just like Google, is a very large collector of information, but unlike Google, the information collected by different surveillance systems is very disparate. On an individual, a State can now easily collect both their banking data, their geolocation over time, the details of their shopping if they pay with an electronic means of payment, all of their communications, and so its social graph in the broad sense, its medical data, etc. Heaps of disparate data that must, to make sense, go through algorithms that we imagine much more sophisticated than those of Google.
The uses that can be expected of all this are very varied. When voting on the Intelligence Law, an amendment, proposed and then withdrawn by the government in the face of the outcry from social networks, envisaged wiretapping all communications as well as the geolocation of the unemployed in order to fight against fraud. It is one use, among others, of Big Data, not necessarily very exciting, but most certainly effective. Recently the Chained Duck reported on a French company, very focused on big data, which on the basis – among other things – of ethnic files extrapolated from places of birth found on electoral files (which are accessible to all), offers politicians cartographic tools intended to assist them in the management of the territory for which they are responsible.
How will such tools evolve tomorrow? It is very difficult to anticipate, and it depends of course on the political will behind it, but also on the world of technology.
Finally, we must not forget one of the holy grails of Big Data which is predictive, that is to say the ability to predict, from a detailed statement of the present state, what will be the future. We can thus imagine predicting what will be your state of health tomorrow: information that can prove to be a fantastic progress for society if it is in the hands of an organization managing public health, or which can be synonymous with great inequalities, if the insurance companies are taking it. The predictive will completely change both the way of managing public health and public order, but also the management of the city as a whole and democracy in general.
If we governed, just yesterday, on the basis of incessant opinion polls, the photograph that we could take of opinion with Big Data and predictive systems should radically change the very idea of democracy and the exercise of power.