Can the Utility of Anonymized Data Be Used for Privacy Breaches?

Group based anonymization is the most widely studied approach for privacy preserving data publishing. This includes k-anonymity, l-diversity, and t-closeness, to name a few. The goal of this paper is to raise a fundamental issue on the privacy exposure of the current group based approach. This has been overlooked in the past. The group based anonymization approach basically hides each individual record behind a group to preserve data privacy. If not properly anonymized, patterns can actually be derived from the published data and be used by the adversary to breach individual privacy.