An Empirical Evaluation of Adaptive Browser Fingerprinting Techniques Against Anti-Fingerprinting Defences

Abstract

Browser fingerprinting has emerged as a powerful tracking mechanism that operates without relying on traditional storage techniques such as cookies. In response, modern privacy-focused browsers claim to mitigate fingerprinting through entropy reduction and isolation mechanisms. However, the effectiveness of these defences against adaptive, real-world fingerprinting techniques remains unclear. This paper presents an empirical evaluation of adaptive browser fingerprinting using FingerprintJS v5 across modern browsers, including Chrome, Brave, and Tor. The study systematically analyses fingerprint stability under various defence configurations, browsing contexts, and identity resets. The outcome of the experiment shows that while privacy-enhancing browsers reduce long-term tracking, adaptive fingerprinting techniques continue to generate stable identifiers within sessions, revealing limitations in current defence mechanisms. The findings highlight the need for stronger and more standardized anti-fingerprinting approaches. Thus, this research has proposed a novel framework that can further enhance the accuracy of adaptive browser fingerprinting techniques.