The idea of using user preferences to assist with information filtering and with providing the most “relevant” answers to queries has recently received some attention from the research community. This has resulted in the proposition of several frameworks for formulating preferences and their direct embedding into relational query languages. In this paper we discuss major exploitation issues and privacy concerns inherent in the basic paradigm used by these proposed approaches when used with e-businesses not implicitly trusted by the user. We then outline an alternative approach geared specifically towards using user preferences when interacting with e-businesses. Categories and Subject Descriptors K.4.1 [Public Policy Issues]: Privacy; H.3.5 [Online Information Systems]: Web-based services; H.3.4 [Systems and Software]: User profiles and alert services General Terms Algorithms, Security Keywords Privacy, User Preferences, E-Commerce, Exploitation