The Limits of Differential Privacy (and its Misuse in Data Release and Machine Learning)

Differential privacy (DP) is a neat privacy definition that can co-exist with\ncertain well-defined data uses in the context of interactive queries. However,\nDP is neither a silver bullet for all privacy problems nor a replacement for\nall previous privacy models. In fact, extreme care should be exercised when\ntrying to extend its use beyond the setting it was designed for. This paper\nreviews the limitations of DP and its misuse for individual data collection,\nindividual data release, and machine learning.\n