Below the transcription of my presentation at the Global Forum 2014. Thanks a lot to the organizers for this great job! See also my after-presentation interview here.
Big data is about knowledge, but what kind of knowledge are we talking about? Some of the experts in the social scientist area are of the opinion that causality doesn’t matter anymore. What matters is correlation between facts, to make analysis less science oriented and more efficiency oriented.
A humorous example of the use of correlation was given, by referring to the exchange between former Marine and combat veteran Joe Pyne and musician Frank Zappa in 1965. It went like this Pyne said, “Well, I see you have long hair. You must be a girl.” Zappa fired back, “So, I see you have a wooden leg. You must be a table.”
Big Data are often considered as enlarging choices and enhancing the equality of choices. However, big (commercial) data is not about social inclusion. Recommendation systems of are a good example: According to your consumer profile, and the products related to a product you’re about to buy, a recommendation system suggests products you might also like. It is sometimes frightening to see how relevant the recommendations are. According to some studies, this system helped Amazon to increase its sales by 20 percent. The problem is that, according to an increasing number of studies, these systems tend to put people in cultural cases, that prevent social mobility.
Big Data, as many emphasise, might be a revolution in the way we produce and understand knowledge. However, the type of knowledge produced by big data is not equivalent to the traditional forms of knowledge production, i.e., science, physics, medicine, etc.
This type of knowledge is peculiar and is particularly oriented, at the very first stage, on an immediate efficiency. It doesn’t really matters anymore why a decision has to be made. It matters that, according to the facts, it has to be made.
Big data is not democracy. Actors who are enabled to make decisions, based on this new type of knowledge, are not the same anymore. There are not necessarily elected or chosen by people. This shift of power can be observed in many examples, if not in our everyday life activities. A very relevant example of how private actors are getting power on issues that used to be the “garden” of public actors is the foodflex programme in the US.
Privacy is not the antidote. We usually see privacy and data protection policies as the answer to draw lines at the potential abuses of information systems and, as we call them now, big data.
Privacy has become quite rapidly obsolete in our information age. Its basic principle states that no data should be gathered about persons without their permission, that should be based on what someone wants to do with the data. In the era of the big data, there is a big contradiction between the very principle of big data and its potential to create innovation, and the very limiting principle of privacy.
This is why the definition of privacy has become a matter of the first importance. The question is: who has the power to define what privacy is? Privacy, as conceptualised by the legislators, is not understood in the same way by citizens. Privacy is a very subjective concept… and a normative one!
To say it provocatively, privacy is not only a tool of protection against surveillance, but also a tool of governance. The more you shape privacy, the more you can control it. A concept that tend to individualize people to make them fit the surveillance projects.
The following question referred to big data algorithms and recommendation systems, like the one used by Amazon. To what extend are people’s choices being eroded? If suddenly it is made so easy to see things, people may miss other things. Or is it bringing greater diversity?
Mr. Coll answered that one has to make a difference between inter-diversity and intra-diversity. It is bringing greater intra-diversity, i.e., diversity only in your own social group. Each social group has a greater diversity than before.