Abstract
Calculating trust in order to give valuable privacy recommendations is an increasing need in pervasive environments, especially when also considering context information (e. g. location information, nearby people). The developed di.me userware should support the user not only by helping in managing the personal information but also by giving valueable recommendations and warnings concerning the (unintended) disclosure of sensitive information to other, possibly untrusted, parties.
In this deliverable we present concepts to measure and establish multidimensional trust metrics, specially tailored to the end-users’ needs in digital.me. Furthermore it includes usability and reliable interaction aspects. Thereby, information obtained from the context could represent a separate dimension in the trust metric. A special focus is the correctness and robustness of a suggested trust metric towards meaningful privacy recommendations for our scenarios.
The deliverable is strongly connected to WP3 (especially task 3.4) and also an important component of the intelligent user interface (WP5), which will present the resulting warnings to the user.
The suggested metric could help to avoid manipulations and attacks based on fundamental requirements for situational trust, by giving meaningful warnings and prove awareness of possible risks to the user.
The metric supports dynamic and (semi-)automatic trust extensions and migration (e.g. using recommended trust settings and preferences).
Among the planned future work is the improvement of trust calculation by taking into consideration presence history of persons and locations, in order to increase the accuracy of our base trust value.
Developed concepts envisage meaningful and strong robustness against manipulations (by intentional or unintentional misusage), in order to reach representative evaluations. The involvement of a large number of end-users implies supporting different trust needs and interpretations, as addressed in section 1.2.7 of the DoW.
About
-
Dissemination level: Public
-
Nature: Report
-
Workpackage: WP 4 Infrastructure for trust, privacy and security