The controversial promises of big data
Our human essence is not in a set of data and algorithms. Despite what the tech gurus sell us, our brains are not easily replicable: no matter how much data they have about your daily behavior, they will never know the real reason for your decisions and will never get to predict your next steps without error.
In the world of Google, Instagram, mobile technologies, cloud computing, big data, artificial intelligence … the data scientists are the trending professionals. Access to massive amounts of data makes us dream, among other things, of predicting and understanding human behavior. To make this a reality we need professionals capable of developing powerful data analysis algorithms, predictive models, and learning algorithms. And mathematicians have a good profile for this type of work. Hence, Mathematics has become the race with the highest marks all over the world.
However, in all this data madness, we forget something fundamental: data is obtained after a phenomenon, but it is not necessarily the cause of the phenomenon. And in many cases what is interesting is the cause.
This is especially true when we talk about understanding human behavior. Today we have infinite data about our behavior: Internet browsing, searches, comments, photos, videos, location, shopping, series, movies, and favorite music, etc. Soon, with the development of the Internet of Things (IoT), we will also have data on our interaction with all kinds of objects — household appliances, lighting, vehicles, clothing, shopping carts, machines vending machines, a job, etc.
In other words, it is quite likely that in the short term we could reproduce with great accuracy the behavior of a specific person throughout a specific day — bypassing the data protection law in force in many countries. However, even in that case, we would still lack information to know why at a certain moment our person X behaves as they behave. And it is that our behavior depends on something more than the environment data.
Brain and computer
In the last fifteen years, this mathematics, convinced of the importance of data and rational decision-making, has been reconverted. In this time many mathematicians dedicated their professional careers to Neuroscience, something that has allowed them to understand that we cannot compare our brain with a computer and that we cannot aspire to discover the algorithm that is capable of predicting our individual behavior, our decisions and our lives.
To begin with, human beings have a big limitation when it comes to perceiving all the data around us.
It is estimated that we are surrounded by more than eleven million bits of information per second, while we can only process fifty every second (Wilson, 2020). This limitation means that, when acting, not all the environmental data influences us, but only those that our brain allows us to process. In this sense, the machines have much greater capacity than us.
In any case, our brain is equipped with some processes — attentional processes — whose objective is to process that information that our brain considers useful at a given moment and discard the rest. Basically, we have two types of attention processes:
Bottom-up processes: exogenous attentional processes where attention is guided to salient stimuli that are relevant due to their properties in relation to the background in which they are found.
Top-down processes: endogenous attentional processes where attention is guided according to previous knowledge, as well as the intentions and objectives of the moment.
If we wanted to understand the behavior of our person X, the first thing we would have to do is determine what information they had previously processed. It would be relatively easy to know the stimuli perceived by the bottom-up process because they are quite similar between humans and even reproducible by an algorithm. However, it would be practically impossible to determine the stimuli perceived by the top-down process since it is totally dependent on the person and operates on the non-conscious plane.
But in addition to all this, human beings do not have an objective way of giving meaning and value to the data we perceive.
As for giving meaning to things, a very clear example was the viral phenomenon #TheDress, a dress that according to those who looked at it saw it as white and gold or blue and black. In this case, it depended on how our brains interpreted the lighting in the photo: “People who correct — unconsciously — for cold [bluer] lighting see the dress in the image as white and for the same reason see the lace as gold . People who correct — also unconsciously — for warm lighting [more yellow or reddish] see the dress as blue and black ”(Brainard and Hurlbert, 2015). But something that became so clear thanks to this viral example happens to us continuously: we interpret the world around us based on the way we have learned that the world is. Actually each person lives a different reality and this reality, their reality, affects their behavior and their decisions (Tversky, 1981).
As for giving value to what we perceive, there are numerous studies that show how human beings are biased when it comes to giving value to things. For example, we tend to value family things more (DellaVigna, 2009), we prefer a reward in the short term although it is much less than a reward in the long term (Laibson, 1997), when something changes, we tend to evaluate the new by the effect of the change (to positive or negative) and not so much for the value itself that it has (the value that it would have in isolation) (Frederick, Kahneman and Mochon, 2010). These biases and many others affect our behavior and decisions. However, they do not affect us in the same way: the effect that these and many other biases have on each person is unique.
So if we were to continue to understand Person X’s behavior, we should control their individual way of perceiving the world and giving it value. Something that is completely out of reach today since they are individual processes, not conscious and not observable.
It is true that in recent years, neuroscience has been making great strides to gain insight into these processes. We have technologies capable of obtaining data directly from the human brain and converting it into emotional and cognitive information. This helps us to know whether, on a non-conscious level, a certain stimulus activates attentional processes or whether it is positively valued by person X. So now we are in a position to add one more link in the chain of understanding of human behavior: “ Our person X behaved at a certain moment as they behaved because they perceived this information and that activated a positive emotion ”. But again: we do not know why person X unconsciously paid attention to this information and perceived it in a positive way.
Never before have we had so many resources to investigate our own human nature
The only thing we know for sure is that the brain is a learning organ: it contains conscious and non-conscious information that is used to guide behavior efficiently and to adapt to the results and effects of behavior. In other words, what really makes us act as we act is a set of learning obtained throughout our existence that makes us the person we are.
Thus, in order to understand and predict the behavior of our person X in a correct way:
• We should have stored the conscious and non-conscious information of its entire existence. However, even if Person X wanted to collaborate and share all the information he has consciously about him, we would still lack access to information about non-conscious processes — which is approximately 80% of brain activity.
• We should be able to understand how the brain uses this information to learn and adapt. However, we are still a long way from understanding how the human brain works in this regard.
So, no matter how much data they have on our behavior, no matter how much Mathematics, prediction models and learning algorithms advance, we will never see machines capable of fully understanding the human essence. I am sure that individual behaviors will continue to be unknown for machines, humans and even for the individual himself.
Now, although it is not possible to accurately predict our human behavior — especially at the individual level — data and technology will certainly offer us a great opportunity to learn about ourselves.
Unfortunately, we are so dazzled by technology that when we talk about the IV Industrial Revolution, we focus almost exclusively on technological advances, where STEM profiles become the protagonists. Meanwhile, Humanities, Art and Social Sciences students must feel like I did a few years ago: like useless romantic fools. We are overlooking the fact that, in the entire history of humanity, we have never had so many resources to investigate about our own human nature, about our decision-making processes, about our emotions, about the search for happiness … Let’s not lose this Opportunity, let humanistic profiles get involved in this revolution too. Only by joining efforts and combining capacities will we be able to face the important challenges that we face with this technological avalanche and make the world a better place for everyone.
References:
Brainard D.H. y Hurlbert A.C. (2015): “Colour Vision: Understanding #TheDress” en Current Biololgy. Junio 29;25 (13).
DellaVigna (2009): “Psychology and Economics: Evidence from the Field” en The Journal of Economic Literature, Vol. 47, №2.
Frederick, S.; Kahneman, D. y Mochon, D. (2010): “Elaborating a simpler theory of anchoring” en Journal of Consumer Psychology, 20(1), 17–19.
Laibson, D. (1997): “Golden eggs and hyperbolic discounting” en Quarterly Journal of Economics, 112, 443–477.
Tversky, K. (1981): ‘The framing of decisions and the psychology of choice’” en Science.
Wilson, T. D. (2002): Strangers to ourselves: Discovering the adaptive unconscious. Cambridge, Belknap Press of Harvard University Press.
