Here are some preliminary thoughts I had on Big-Data as it pertains to human behavior.
The utility of Big-Data lies in its ability to predict the behavior of humans. But human behavior is affected by what we know: when my basketball coach tells me I am not bending my knees on my free throw, I change my behavior given this new knowledge. I bend my knees the next time around.
So what happens when we are given access to Big-Data, access to our own behaviors? Well, we see this play out in money-tracking apps. Because my overall spending for a month easily escapes my grasp, which is restricted to more day-to-day memories, seeing my total expenditure for the month is almost always surprising. I spend a lot. I then decide, on this basis, to spend less.
But, as I’ve said, the utility of Big-Data is its ability to predict behavior, and it seems at first that the money-tracking app is a good example of this. Yet, if I gave my app data to someone else while allowing myself to access it too, the other person cannot reliably predict my behavior, at least without knowing me personally. They must know how I would react to the data, and how I would react to the data about my reaction to the data, and how I would… The recursive interplay between my behavior and my knowledge of my own behavior generates an infinite pattern that is un-graspable, unpredictable. For that other person to successful predict my behavior using this app, I must not have access to it, and therefore not be able to change my behavior according to the data I see about myself. The same is true of Big-Data.
What I’m trying to say here is that Big-Data’s usefulness relies on the separation between the ones doing the predicting and the ones being predicted. This is why YouTube, Google, and all these companies cannot ever release details about the data they collect and the algorithms they use to make sense of it, because as soon as they do so, its (i.e. Big-Data’s) value is lost. It can no longer predict. In other words, Big-Data is predicated on power and domination. It is always asymmetric. It is not about how to implement it correctly.
If we are truly living in the age of Big-Data, as many have claimed, then really, what’s all the fuss about? After all, power and domination are nothing new.