Sometimes called “the internet of things,” Big Data has arrived. It will “replace ideas, paradigms, organizations and ways of thinking about the world,” said Professor Brynjolfsson, Director of M.I.T.’s Center for Digital Business at a recent conference. Well, maybe. But it’s worth thinking about what it might not be able to do.
As Steve Lohr put it in his year-end review of the field in The New York Times, such claims rely “on the premise that data like Web-browsing trails, sensor signals, GPS tracking, and social network messages will open the door to measuring and monitoring people and machines as never before.” Computer algorithms, using that data, will enable us to “predict behavior of all kinds: shopping, dating and voting, for example.”
All that is true, and we see this at work as the internet already tracks every search we make on our computers. We can’t escape innumerable hints and suggestions of what else we might want to buy. Nothing is forgotten or ignored. And those are the more detectable signs of how we are being tracked.
But as Lohr points out such predictions are based on mathematical models and our models are made by human intelligence. Once set up, the models crunch data quickly and efficiently, but, being devised by humans, they themselves are not only fallible but also vulnerable to misuse.
Much attention has been paid to the invasion of privacy inherent in such models. What are we inadvertently revealing about ourselves? And who will use that information to manipulate and control us? It is happening now, of course, but it will only get worse. And how will we know?
A danger of another kind is the lack of sophistication and accuracy in the models used. Good programs require math and computer skills but also an ability to be innovative and thoughtful. Lohr notes that the McKinsey Global Institute projected that the U.S. would need 140,000 to 190,000 more workers with “deep analytic expertise.” He quotes Claudia Perlich, chief scientist at an online adtargeting startup in New York: “We can’t grow the skills fast enough.”
It is not just computer and math skills that are needed. Lohr notes: “Listening to the data is important, but so is experience and intuition. After all, what is intuition at its best but large amounts of data of all kinds filtered through a human brain rather than a math model?” (See, “Sure, Big Data Is Great. But So Is Intuition.”)
To be clear, that includes the unconscious information to which we are inattentive because it sometimes seems irrelevant, sometimes unfashionable and sometimes unwanted. The point is that it is often precisely that information — unsought, unexpected, perhaps even difficult to accept or grasp – that reveals what we most often need to know.
At M.I.T.’s recent conference, Lohr reported that a panel asked about big failures in Big Data could come up with no examples. Later, however, someone in the audience commented that Big Data failed to foretell the credit crisis and financial crash of 2008. Oh!
Could it be that the specter of its potential leads its adherents to neglect or downplay the human factor? Does Big Data make people over-confident or smug? If so, that’s just the kind of problem that Big Data can’t solve?