KHGI 2005 Poster

							      Managing the Privacy of Incidental Information During Collaboration
                                                                   Kirstie Hawkey and Kori Inkpen
                       Concept                                                 Dalhousie University                                                                        Challenges
                                                                            {hawkey, inkpen}@cs.dal.ca
          Ad hoc                        Incidental
                                                                                                                                                           Complexity Reduction
       Collaboration                   Information                                                                                                          Large volume of information.
   Colleagues often gather          Many traces of past
                                                                                 Privacy Patterns                                                           Multiple contexts of creation and viewing of incidental
     around a computer            activities are visible with        Patterns on a per-window basis suggest a                                                information.
                                     casual inspection.              semi-automated approach to privacy                                                     Must balance amount of control with maintenance time
                                                                     management may be feasible:                                                             and effort.
   Individual                                    Privacy                                                                                                    Privacy patterns suggest a semi-automated approach
  Behaviours                                                                                                              Rapid Bursts of                    is feasible.
                                            Normative privacy
                                                                                                                          Browsing
 Privacy concerns                              for personal                                                                                                Data Analyses
   and personal                             displays does not                                                                                               High volume of logged data during field studies.
    information                            apply; the display is                                                                                            Numerical averages insufficient due to individual
                                                                                                                          Large Volume of
management styles                            an object in the                                                                                                differences.
                                                                                                                          Pages
are very individual. Personal Information     collaboration.                                                                                                Data mining and visualization techniques may uncover
                        Management                                                                                                                           patterns.
                                                                                                    Sequential
                                                                                                    Representation
                                                                                                                             Streaks                       System Evaluation
                  Managing large amounts of
                                                                                                                                                            Field setting required for evaluation of management
                   personal information is
                                                                                                                                                             solutions.
                         complex.
                                                                                                                                                            To encourage natural behaviour, users need to
                                                                                                                                                             maintain privacy.
                                                                                                                             Transitions                    Large volume make fine-grained self-reports tedious.
    Field Study: Privacy Gradients
                                                                                                                   Temporal
                                                                                                                   Representation                                             Next Steps
                                                                                                                                           Temporal
                                                                                                                                           Patterns        Potential Solutions
                                                                                                                                                            Browser windows of differing privacy levels?
                                                                                                                                           Window           Filter which incidental information is displayed.
                                                                                                                                           Re-visitation    Classify new information as it is generated.
                                                                                                                                                           Research studies
Methodology                     Goodness of Fit                       Clusters                               C1      C2      C3      C4
                                                                                                                                           Overall          Survey of privacy concerns with respect to incidental
                                                                      Privacy Gradient             Overall        Final Cluster Centers

 Field study for one week.      75% of gradients fit most                              Public       42% 22%        36%     62%     18%   Privacy           information (underway).
 20 laptop users (multiple       of the time; 20% fit all of
                                                                                  Semi-Public         25% 58%        21%     16%     28%
                                                                                                                                           Gradient         Field study examining privacy in context of location
                                                                                         Private      15% 9%         36%     11%     9%

  contexts).                      the time                                         Don’t Save         18% 11%        7%      11%     46%   Usage             and pages (underway).
 Client-side logging.           Hard to classify:
                                                                      Number of Participants                 3       5       10      2
                                                                                                                                           Patterns         Longitudinal field study of management solution
 Electronic diary.               Sites with multiple
 Privacy gradients.                purposes
                                                                                                                                                                          Acknowledgements
 Paper classification tasks.     Sites with variable
                                                                                                                                                                 Thanks to the members of the EDGE Lab
                                    content
                                                                                                                                                                    for their support. This research is
                                                                                                                                                                         funded in part by NSERC