Predicting User Privacy Preferences based on Dynamic Interpersonal Relationships and Content Sensitivity Analysis

Users are increasingly sharing more self-generated content online. Such content can however endanger users’ privacy and have serious consequences if shared to an inappropriate audience. The current state-of-art to manage the audience of the shared content has repeatedly been demonstrated to be inefficient in appropriately supporting users in this task. In this project, we therefore explore a new approach to assist users in selecting the audience of their content. Our proposal aims at leveraging both the sensitivity of the content to be shared as well as the relationship of the user with the intended audience to make suggestions to users. The ultimate goal of our solution is hence to allow users to simultaneously benefit from sharing content online while better protecting their privacy in a more usable fashion.


  • L. Kqiku and D. Reinhardt. Lengthy Early Morning Instant Messages Reveal More Than You Think: Analysing Interpersonal Relationships using Mobile Communication Metadata. Pervasive and Mobile Computing (PMC), 2023. [PDF]
  • A. Tillmann, L. Kqiku, D. Reinhardt, C. Weisser, B. Säfken, and T. Kneib. Privacy Estimation on Twitter: Modelling the Effect of Latent Topics on Privacy by Integrating XGBoost, Topic and Generalized Additive Models. Proceedings of the 8th IEEE International Conference on Privacy Computing (PriComp), 2022. Accepted for publication. Acceptance rate for full papers: 23%.
  • L. Kqiku, J. Dieterle, and D. Reinhardt. Exploration of a Mobile Design for a Privacy Assistant to Help Users in Sharing Content in Online Social Networks. Proceedings of the 8. Usable Security und Privacy Workshop (MuC workshop), 2022. [LINK]
  • L. Kqiku, M. Kühn, and D. Reinhardt. From Sentiment to Sensitivity: The Role of Emotions on Privacy Exposure in Twitter. Proceedings of the 2nd ACM Workshop on Open Challenges in Online Social Networks (OASIS, HT workshop), 2022. [PDF]