IEEE Future Directions
Related Articles
I did attend the kick off meeting for the IEEE BiG Data Initiative at Stevens, NJ, and I enjoyed the insightful presentations offered. I also enjoyed the following discussion. Among several other things it made me think that we are possibly on the edges of a second revolution in the path to dematerialisation.
Now, I know that there have been several claims of revolutions on the horizon for Big Data and several (understandable) confutations to those claims.
My thoughts are not in the field of how BiG Data could revolutionise our way of living or doing biz, rather on the impact that Big Data can have on values, both economic and ethical. And this is what I shared there and I am sharing with you now.
We have had a first revolution when we started to make use of digital "stuff" replacing atoms. Consider the music industry: moving from vinyl records to mag tape did not changed the music industry, it was a shift from an "atom" support to a different "atom" support. However, when the shift was from "atom" to "bits" through Mp3 players the music industry took a blow (micro blows were given also in the shift from vinyl to tapes because of the easiness of copying tapes) that changed the biz completely. This is also seen in paper publishing, in movies and so on.
The fundamental reason is that while the cost in duplicating an object made of atoms requires expenditure to get more atoms and manufacture them, in the case of bits there is no cost involved since bits come for free, and the cost of copying and dispatching bits is negligible. A bit copy is as good as the original but doesn’t cost a penny. In the bit space you don’t need the original, a copy is good enough, actually, it is indistinguishable from the original. It took, and it is taking, various industries a lot of effort to reinvent the ownership concept in the bit area, with names like licences, creative commons….
Now consider BiG Data. Here you have some sort of magic happening. You have bits that are owned by somebody, that might be protected by some sort of licences so that ownership is rewarded. And now through Big Data Analytics, but also through BiG Data visualisation you can create new information having in many cases a value that is much greater than the one of the original data. Who is the owner of such value? Are the owners of the data you have been used to create the brand new information entitled to sharing the value produced? May be but at the same time it will be extremely difficult to assess the appropriate "reward". We need a new biz framework in place, and new technologies to make it possibile.
There is also another, disturbing, twist. With ownership also comes accountability. But in this new world of Big Data where ownership is fuzzy also accountability is fuzzy. Think about privacy (you can see this in terms of ownership of your privacy).
You are walking in a street and your image is captured by a camera. It was your choice to walk on that street at that particular time so there is no big deal. But now consider that the image captured by that camera becomes part of the Big Data in that city. A Big Data analytics can work out your identity by correlating that image with the image taken by the safety camera pointing at your home. Well, no problem there since by walking in a street you accept that somebody will know you and will see you. But now consider that in the Big Data space it is possibile to correlate your image on that day with images of you taken in the same point in previous days. Out of this a pattern can emerge. This pattern can be cross referenced with a pattern of another person in a different part of the city that at a slightly different time, let’s say ten minutes before the time your image was taken, generate a similar pattern. And then a further correlation shows that you and that person are converging onto a location. Out of that one can see the information that the two of you are related and are meeting for an hour at a specific place (or even at different places). This is an emerging information that was basically impossibile to detect (at least at a low cost) in the past. And this is going to have disruptive consequences.
The topic of emerging information is a new one, particularly associated to the low cost of this information. And this, in my mind, is creating a second revolution: the first one was once we moved from "atoms" to "bits", the second one as we are moving from "bits" to the creation of new "bits" through correlation.
