Given that no one AFPET evaluation can match all goals, we will explore points in the space of possible evaluations while focusing more on prospective evaluations. If instead the goal is to determine which AFPET to fund for further development, the user numbers of today may matter less than the technical or theoretical capabilities of the AFPET. For example, if the goal of evaluation is to determine which AFPET to use today, one would want to know how many other users of the AFPET there are since they will form the anonymity set – the group of other users one will blend in with. We develop a method that compares the trackability of AFPET-modified fingerprints with those of the original fingerprints.ĭepending on the goals, AFPET evaluation could depend on the context in which the AFPET is used, accounting for features of other users and non-users, or be a more theoretical assessment of the AFPET's potential, untied to the vagaries of today. Our goal is to find attributes that AFPETs are not masking and to quantify their effects on privacy. Studies have also uncovered fingerprinting code on popular webpages .Īnti-Fingerprinting Privacy Enhancing Technologies (AFPETs) aim to protect consumers against fingerprinting by masking, or spoofing, the values of attributes. The list of attributes that can be used in fingerprints is rapidly increasing . Fingerprints collected on websites like and /fp demonstrate that they are sufficiently unique and stable for tracking purposes . A stateless tracker extracts fingerprints from consumers as a collection of several attributes of the browser, operating system, and hardware, typically accessed through Javascript APIs. In addition to using stateful methods, such as cookies, some trackers use stateless tracking mechanisms, also known as browser fingerprinting.

Online data aggregators track consumer activities on the Internet to build behavioral profiles.

Canvas defender for firefox 45.0