Predictive Analytics, Revised and Updated

Chapter 2With Power Comes ResponsibilityHewlett-Packard, Target, the Cops, and the NSA Deduce Your Secrets

How do we safely harness a predictive machine that can foresee job resignation, pregnancy, and crime? Are civil liberties at risk? Why does one leading health insurance company predict policyholder death? Two extended sidebars explore: (1) Does the government undertake fraud detection more for its citizens or for self-preservation, and (2) for what compelling purpose does the National Security Agency (NSA) need your data even if you have no connection to crime whatsoever, and can the agency use machine learning supercomputers to fight terrorism without endangering human rights?

Predictive analytics…is right at the fulcrum point of utopian and dystopian visions of the future.

–Andrew Frank, Research Vice President, Gartner

What would happen if your boss were notified that you're allegedly going to quit—even though you had said this to no one? If you are one of the more than 300,000 who work at Hewlett-Packard (HP), your employer has tagged you—and all your colleagues—with a “Flight Risk” score. This simple number foretells whether you're likely to leave your job. As an HP employee, there's a good chance you didn't already know that. Postpone freaking out until you finish reading the full explanation in this chapter.

This story about HP arrived in the wake of media outcry ...