The real danger in the U.S. electronic surveillance programmes revealed by Edward Snowden is that governments might use the data to punish people not for what they do, but for what they might do, says Viktor Mayer-Schönberger, one of the world’s leading data experts on big data.
“The problem is not big data, but how big data is used,” Mayer-Schönberger said. “The real danger is that it creates algorithms that predict human behavior – how we will behave in the future from how we (or other people) behaved in the past … and penalizes us for something we might do rather than for something we have done.”
Mayer-Schönberger, Professor of Internet Governance and Regulation at the University of Oxford’s Oxford Internet Institute, noted that the U.S. government says it is using the surveillance to prevent terrorist attacks before they happen. While the ability to prevent a crime might sound like a good thing, Mayer-Schönberger said that it comes with an inherent weakness: data reveal a probability, not a certainty.
“By punishing people for the probability that they will commit a crime, you deny them their humanity – their free will to decide when and when not to act,” he said. “It’s punishment without proof … [because] we would never know if they would commit the crime or if they would not.”
Mayer-Schönberger’s comments came at a JMSC-sponsored talk at the University of Hong Kong on June 20. The theme of the talk was his new book, “Big Data: A Revolution That Will Transform How We Live, Work, and Think,” but the recent leaks of information about the U.S. programs by Edward Snowden, a former contract employee at the National Security Agency, naturally came up.
Mayer-Schönberger said that the NSA secrets leaked by Snowden illustrate one of the dangers of big data: by removing free will from the commission of a crime, human responsibility is also removed.
“If I am predicted to commit a crime, and going to be punished for that prediction, then why wouldn’t I commit it?” he said. “By endangering free will, we may turn our society into one without responsibility. Because without guilt then there can be no innocence.”
To prevent the abuse of the government’s data-gathering ability, Mayer-Schönberger recommended four things: new laws to hold people accountable for how they use data; safeguards to free will similar to guarantees for procedural fairness; a new class of professionals, called “algorithmists,” who understand the complexity of data and take an oath of impartiality; and regulations to prevent the rise of “big data barons,” ensuring data is not held in the hands of fewer and fewer people.
“We need to carve out space for our reason and imagination, and for acting in defiance of what the data tells us,” he said. “Because data is always just a shadow of reality and therefore always imperfect and incomplete. As we walk into the big data age we need to do so with humility and humanity.”