Impacts of user privacy preferences on personalized systems: a comparative study

Abstract

Recent developments in privacy awareness and legislation may have a significant impact on the advancement of personalized systems. Though many countries have enacted comprehensive privacy laws, user concerns are still high. We compared 30 opinion surveys on Internet privacy, categorized the responses, and matched them with possible impacts on personalized systems. The analysis of a cross-section of privacy surveys should provide a more objective view of consumer concerns than results from a single study. This research thus represents a first contribution towards the identification of requirements for privacy-preserving personalization, to improve users’ trust when interacting with personalized systems. Introduction Personalized (or “user-adaptive”) systems have become increasingly popular since the beginning of the 1990’s, and have gained substantial momentum with the rise of the World Wide Web. The market research firm Jupiter defines personalization as predictive analysis of consumer data used to adapt targeted media, advertising, or merchandising to consumer needs (Foster 2000). A more general definition was proposed by Kobsa et al. (2001) who regard a personalized hypermedia application as a hypermedia system which adapts the content, structure and/or presentation of the networked hypermedia objects to each individual user’s characteristics, usage behavior and/or usage environment. In contrast to user-adaptable systems where the user is in control of the initiation, proposal, selection and production of the adaptation, user-adaptive systems perform all steps autonomously. A well-known example of a personalized website is Amazon.com, which generates purchase recommendations based on a user’s purchase and interaction history. Other examples of personalized web sites are listed in (Dean 2000). A categorization of user-adaptive systems according to Kobsa et al. (2001) is depicted in Table 1. The advantages of personalization can be manifold. Online users see the major benefits in sites being able to offer more relevant content and to recall user preferences and interests (Cyber Dialogue 2000). However, personalization of hypermedia presentation is beneficial for several other purposes as well, most notably for improving the learning progress in educational software (Brusilovsky et al., 1998; Eklund & Brusilovsky, 1998; Specht, 1998). In order to deal with impacts of privacy concerns on personalization systematically, it is helpful to identify privacy-critical personalization processes. Personalization can basically be depicted as a cycle of recurring processes consisting of data collection, profiling and matching (Foster 2000). From collected data, user profiles can be created that are the basis for adapting user interfaces to individuals or groups of individuals. We focused on the step of data collection because it is the most privacy-critical in the personalization process. The collection of extensive knowledge about users’ interests, behavior, demographics and actions is necessary for most user-adaptive systems. However, this could provoke privacy fears that limit consumers’ willingness to share information. We looked at privacy issues from a user’s standpoint and not from a legal point of view. While privacy legislation should serve and protect the users’ interests, user concerns should be a central starting point for the legislative process. For a recent discussion of privacy legislation in different countries we refer to the International Survey of Privacy Laws and Developments (Electronic Privacy Information Center 2002). Data Sources The amount of personal data that is available online has rapidly increased over the years. Individuals often transmit personal information online, either actively by submitting data (e.g. a shipping address for books), or passively, by leaving electronic traces in log files both at the server side as well as in the network. Improved accessibility of data – not only from the World Wide Web but also from multiple user touch points and external data sources – have further increased the amount of information available about individuals. 1 This does not seem to be very much to be the case at the moment: studies suggest that legislative actions seem to only have a marginal impact on consumer concerns (EU/Interactive Policy Making, 2002) Kobsa (2001) partitions data types into user data, usage data, and environment data. User data denotes information about personal characteristics of the user, while usage data is related to a user’s (interactive) behavior. Usage regularities are based on frequently re-occuring interactions of users. Environment data focuses on the user’s software and hardware and the characteristics of the user’s current locale. Personalization systems often need to acquire a certain amount of data before they can start adapting to the user. Thus, they are often only useful in domains where users engage in extended (and most often repeated) sessions of system use. They may not be appropriate for infrequent users with typically short session. No. Input Data of UserAdaptive Systems Examples of User-Adaptive Systems

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 94 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.