By Lloyd Marino
The questions and criticisms were bound to come up. “The backlash against big data,” proclaimed a recently published Economist article. “Eight (No, Nine!) Problems With Big Data,” shouted a New York Times opinion piece. “Big data: are we making a big mistake,” questioned the venerable Financial Times. A piece in Spiked went even further, boldly suggesting that Big Data may be “squishing our humanity,” and mocking the idea of a “brave new world [where] people become passive data points. It is their data that is the active ingredient. Big Data… now [has] the capacity to reveal things to us about our daily lives we weren’t even conscious of. Algorithms will set you free, brothers and sisters.”
These are just a handful of many articles that have openly challenged all of the hullaballoo surrounding Big Data and even called its value into question. What gives? Why is the information explosion suddenly being challenged on a number of fronts? More importantly, are these observations truly accurate? Is the criticism fair? I’d pose a different question. Is Big Data “squishing our humanity,” or simply revealing something more fundamental about us? Indeed, far from “squishing” anything, I’d argue that Big Data sheds light on what it means to be human in a technology driven world.
The Economist piece, as an example, points out what it sees as three major problem areas for Big Data and data analysis: (1) human bias, (2) the need for subject-specific expertise to create specialized algorithms based on data and (3) the issue of “spurious correlations,” which is the idea that two variables that have no direct causal relationship may be wrongly inferred that they do. (Imagine if I held up a graph showing the rates of marital infidelity and Oreo cookie consumption tracking each other closely over a decade. Tempted to believe there’s a link? Perhaps there’s a chemical in the cookie that leads people astray?)
Well, OK. While these are issues, even valid criticisms perhaps, what do the naysayers expect? Let’s take the issue of human bias. If we learned anything from quantum theory, we know that the very act of watching affects the observed reality. A study conducted at Israel’s prestigious Weizmann Institute of Science found that simply observing a beam of electrons affects the way the beam behaves. The experiment revealed that the greater the amount of “watching,” the greater the observer’s influence on what actually takes place. I should add, however, that while quantum theory may change our understanding of submicroscopic activity, it doesn’t invalidate Newtonian laws. If you push something hard enough, it still falls over. Jump out of a second story window without a net to land in, and you’ll still hit the ground, as far as I know.
This makes absolute sense, of course. All studies are inherently biased. Researchers venture into their labs wanting, even expecting certain results. They set up the experiments, pick the controls, and measure outcomes. We think of the scientific method as being objective and rigorous in separating out fact from fiction, or more precisely, what is true from what we wish were true. And in some ways, it is all of those things. However, if you’ve ever been in a science class, you’ll know that it’s easy to manipulate results, deliberately, unintentionally or unconsciously. At every step of the scientific method, there’s room error, bias, and outright distortion.
My comments aren’t intended as a blanket indictment of science or the scientific method. I’m deeply indebted to the important contributions science makes on a daily basis to the advancement of human civilization, and life on this planet in general. But in any endeavor, you can’t completely remove bias from the equation.
As it relates to Big Data, I think there’s confusion between data and information imbued with knowledge and meaning. Human experience arbitrates between information and meaning. Think about it. Information, or data, when you get right down to it, is just stuff, an undifferentiated stream of raw material that has to be interpreted and analyzed before arriving at meaning. And this is where the human element comes into play. No matter how experienced we become at interpreting and analyzing data, that doesn’t mean we’re going to draw accurate conclusions. People make mistakes, boat loads. It’s part of being human. That’s not going to change, no matter how sophisticated we become. Remember, we’re the ones who created and implemented all of this great technology. We created Big Data. It’s reasonable to conclude that Big Data—from collecting to interpreting—is the very embodiment of what it means to be human. The information that comes out of Big Data generated algorithms can offer suggestions, even point us in certain directions. However, we must ultimately decide what we do with that information moving forward.
Big Data will never do away with human subjectivity, which is why questions about human bias, while understandable, are misplaced, redundant even. No matter how sophisticated our methods for collecting, analyzing and interpreting information become, at the end of the day, there’ll still be a collection of primates tasked with making sense of all of it. If we want Big Data and all its possibilities, then we’re going to have to embrace its limitations. The bigger challenge I’d argue is to ensure that every attempt is made to use Big Data to advance civilization, and move humanity forward in positive fashion. While, as a Socrates said, an “unexamined life is not worth living,” perhaps given all we know about ourselves, an over-examined life is no day at the beach either.