From buying books to finding the perfect partner, we share our most intimatewants and needs with our favourite online systems. But how far should we acceptpromises of privacy in the face of personal profiling? In particular we ask howcan we improve detection of sensitive topic profiling by online systems? Wepropose a definition of privacy disclosure we call{\epsilon}-indistinguishability from which we construct scalable, practicaltools to assess an adversaries learning potential. We demonstrate our resultsusing openly available resources, detecting a learning rate in excess of 98%for a range of sensitive topics during our experiments.
展开▼