Privacy decision-making

Helping people to make decisions about their privacy-sensitive data.

The Internet is increasingly sprawling with social networks that allow users to share a variety of content with each other, and personalized services that provide users with content tailored to their context and personal preferences. However, not all users are comfortable providing their private information to these systems. They thus have to make a trade-off between possible privacy threat of providing the data and the benefit they may incur when the system can use the data, a process that has been dubbed privacy calculus. The goal of this project is to help users with this decision process.

Defaults and justifications

To effectively make privacy trade-offs, users need take control over their information disclosure decisions, which can be a burden that they may try to avoid. For example, in one study I found that form auto-completion tools (which have become a standard feature in most modern web browsers) make it so easy to submit a fully completed form that users seem to skip the privacy calculus altogether. We developed two auto-completion tools that make it easier for users to control their disclosure.

It is easy to go overboard on control, though: Facebook, for instance, has resorted to “labyrinthian” privacy controls to satisfy the wildly varying privacy preferences of their user base. As a result most Facebook users do not even seem to know the implications of their own privacy settings! In my thesis I try to solve these problems by analysing users' decisions and then giving them smart default settings. These defaults are tailored to the specific user and the specific situation in which the decision is being made. This personalized solution is simplfied by some of the regularities in users' behaviors that we have discovered in our research: there are several broad dimensions of private information, and there seems to be a limited set of distinct privacy profiles of people that seem to disclose similarly along these dimensions.

Another solution is to give users good justifications for providing (or not providing) the system their personal information. The content of such justifications can range from explaining the reason for the request (e.g. how the system would use the data), the benefits of disclosure (e.g. how much better the recommendations of a recommender system will be), or an appeal to the social norm (e.g. how many other users disclosed the information).

Interestingly, though, it seems that these justifications do not always have the intended effect. Although users in our online study felt that the justification messages were helpful, the messages actually decreased their overall trust and satisfaction, as well as their level of disclosure! This may happen because the system disappoints the user with the justifications, as they will at times inevitably predict low benefits. Another explanation could be that the mere presence of a justification sensitizes users to privacy-related issues. Does this mean that justifications just do not work? It could be that I just have not found the "ultimate" justification yet. Another option is that these justifications work for some users, but not for others. We explore this option in our IUI paper.

Current activities



Wisniewski, P., Knijnenburg, B.P., Richter Lipford, H.: Making privacy personal: Characterizing social network users by their privacy proficiency and management strategies. International Journal of Human-Computer Studies (IJHCS), download here.

Dong, C., Jin, H., Knijnenburg, B.P.: PPM: A Privacy Prediction Model for Online Social Networks. Poster paper at the International Conference on Social Informatics (SocInfo), download here.

Knijnenburg, B.P., Cherry, D.: Comics as a Medium for Privacy Notices. Paper presented at the SOUPS2016 Workshop on The Future of Privacy Notices, download here.

Bidgoli, M., Knijnenburg, B.P., Grossklags, J.: When Cybercrimes Strike Undergraduates. Full paper at the Symposium on Electronic Crime Research (eCrime), download here.

This paper won the best paper award (among 11 accepted papers)

Kobsa, A., Cho, H., Knijnenburg, B.P.: The Effect of Personalization Provider Characteristics on Privacy Attitudes and Behaviors: An Elaboration Likelihood Model Approach. Journal of the Association for Information Science and Technology (JASIST), DOI: 10.1002/asi.23629, download here.

Knijnenburg, B.P., Bulgurcu, B.: Form Auto-completion Tools Designed for Elaboration: Overcoming the Deleterious Effects of Decisional Heuristics on Users' Privacy. Paper accepted to the 2015 Dewald Roode Information Security Workshop, available upon request.

Knijnenburg, B.P.: A User-Tailored Approach to Privacy Decision Support. PhD here.

Li, Y., Knijnenburg, B.P., Kobsa, A., and Nguyen, M-H.C. (2015): Cross-Cultural Privacy Prediction. Paper presented at the SOUPS2015 2nd Annual Privacy Personas and Segmentation (PPS), available upon request.

Dong, C., Jin, H., Knijnenburg B.P.: Predicting Privacy Behavior on Online Social Networks. Full paper with oral presentation at the International AAAI Conference on Weblogs and Social Media (ICWSM) 2015, download here.

Acceptance rate: 19%

Wisniewski, P., Islam, N., Knijnenburg, B.P., Patil, S.: Give Social Network Users the Privacy They Want. Full paper at the ACM conference on Computer Supported Cooperative Work (CSCW) 2015, DOI: 10.1145/2675133.2675256, download here.

Acceptance rate: 28%

Knijnenburg, B.P., Kobsa, A.: Increasing Sharing Tendency Without Reducing Satisfaction: Finding the Best Privacy-Settings User Interface for Social Networks. Full paper at the International Conference on Information Systems (ICIS) 2014, download here.

Knijnenburg, B.P.: Information Disclosure Profiles for Segmentation and Recommendation. Paper at the SOUPS2014 Workshop on Privacy Personas and Segmentation (PPS) 2014, download here.

Wisniewski, P., Knijnenburg, B.P., Richter Lipford, H.: Profiling Facebook Users' Privacy Behaviors. Paper at the SOUPS2014 Workshop on Privacy Personas and Segmentation (PPS) 2014, download here.

Wu, H., Knijnenburg, B.P., Kobsa, A.: Improving the prediction of users’ disclosure behavior... by making them disclose more predictably? Paper at the SOUPS2014 Workshop on Privacy Personas and Segmentation (PPS) 2014, download here.

Kobsa, A., Knijnenburg, B.P., Livshitz, B.: Let’s Rather Do It at My Place? Attitudinal and Behavioral Study of Privacy in Client-Side Personalization. Full paper at the ACM SIGCHI Conference on Human factors in computing systems (CHI) 2014, DOI: 10.1145/2556288.2557102, download here.

Acceptance rate: 23%

Knijnenburg, B.P., Kobsa, A., Jin, H.: Counteracting the Negative Effect of Form Auto-completion on the Privacy Calculus. Full paper at the International Conference on Information Systems (ICIS) 2013, download here.

Acceptance rate: 26%; also presented as a poster at SOUPS 2014

Knijnenburg, B.P., Kobsa, A., Jin, H.: Dimensionality of information disclosure behavior. International Journal of Human-Computer Studies (IJHCS), DOI: 10.1016/j.ijhcs.2013.06.003, download here.

IJHCS is ranked #1 among 26 HCI journals (Microsoft Academic Search)

Knijnenburg, B.P., Kobsa, A.: Helping Users with Information Disclosure Decisions: Potential for Adaptation. Full paper at the conference on Intelligent User Interfaces (IUI) 2013, DOI: 10.1145/2449396.2449448, download here.

Acceptance rate: 22%

Knijnenburg, B.P., Kobsa, A: Making Decisions about Privacy: Information Disclosure in Context-Aware Recommender Systems. ACM Transactions on Interactive Intelligent Systems (TiiS), DOI: 10.1145/2499670, download here.

Knijnenburg, B.P.: On the dimensionality of information disclosure behavior in social networks. Position paper at the CSCW2013 workshop on Measuring Networked Privacy, download here.

Knijnenburg, B.P., Kobsa, A., Saldamli, G: Privacy in Mobile Personalized Systems: The Effect of Disclosure Justifications. Short paper at the SOUPS 2012 Workshop on Usable Privacy & Security for Mobile Devices (U-PriSM) 2012, download here.

Knijnenburg, B.P., Kobsa, A., Moritz, S., Svensson, M.: Exploring the Effects of Feed-forward and Feedback on Information Disclosure and User Experience in a Context-Aware Recommender System. Full paper at the UMAP 2011 Workshop on Decision Making and Recommendation Acceptance Issues in Recommender Systems (DEMRA) 2011, download here.

Other activities

Program committee member of PACIS 2016, special track on User-Centered Decision Support Systems in the IoT Era.

Program committee member of the International Conference on Financial Cryptography and Data Security (FC) 2016.

Program committee member of the International Conference on HCI in Business (HCIB) 2016.

Program committee member of the iConference 2016.

Program committee member of the Privacy Enhancing Technologies Symposium (PETS) 2015 and 2016.

Associate Editor for the European Conference on Information Systems (ECIS) 2015.

Student volunteer co-chair of the iConference 2015.

Keynote speaker of the workshop on Human Decision Making in Recommender Systems (Decisions@Recsys2013), read the 2-page abstract here.

Keynote speaker at the CSCW 2013 workshop on Measuring Networked Privacy".

Editorial Review Board member of the special issue on Reframing Privacy for the International Systems Journal (ISJ).

Program committee member of the International AAAI Conference on Weblogs and Social Media (ICWSM) 2013.