The Dangers of a Data-Driven World
As we increasingly use algorithms that base decisions on past successes, innovation is threatened.
BY David Sirota
If the recent political era has taught us anything, it has reiterated the enduring truth of George Santayana's aphorism about memory and duplication. Whether once again watching tax cuts fail to deliver a promised economic boost or witnessing more wars fail to deliver stability, we are reminded that “those who cannot remember the past are condemned to repeat it.”
But then, as much as those haunting words are meant as a warning, technology is today coding Santayana's principle into society's operating system, as if mimicking history is an admirable objective. Indeed, whether it’s movie studios, record companies, government intelligence agencies or corporate human resources departments, algorithms that use the past to predict—and create—the future are making more and more decisions.
For those employed in creative endeavors, it's comforting to believe that technology's use in the information economy begins and ends with the kind of straightforward processes (data entry, dictation, etc.) that require little cognitive analysis and even less artistic thinking. Yet, as Christopher Steiner shows in his mind-blowing new book Automate This, algorithms taking into account past commercial successes are being deployed by the film and music industries to choose which movie and album proposals will be produced. What's more, an increasing number of the algorithms' selections have proven profitable.
Steiner also documents the CIA’s seeming preparation for a real-life version of the WOPR from the 1980s flick War Games. Through grants to New York University and the Hoover Institution, the agency is trying to algorithmically quantify the history of past political and military decisions for the purpose of predicting—and perhaps eventually shaping—future events. ................(more)