Poster: FP-Agent: Towards Fingerprinting AI Browsing Agents

AI browsing agents are an emerging class of AI-driven bots capable of autonomously navigating websites. Unlike traditional web bots, AI browsing agents typically operate using real browsers and perform everyday tasks, making them difficult to detect. In this poster, we present FP-Agent, the first controlled measurement study of seven AI browsing agents and human users. To evaluate the effectiveness of existing industrial bot detection solutions, we first perform a case study, finding that only one of the seven AI browsing agents are detected. Motivated by these results, we use an instrumented honey website to collect the browser and behavioral fingerprint features produced by AI browsing agents and humans while performing three tasks modeled after common real-world interactions: flight booking, online shopping, and forum interaction. We then train a multi-class classifier to analyze the discriminative power of these fingerprint features. We find that AI browsing agents can have identical browser fingerprints with each other and with humans. Behavioral fingerprints, however, are distinctive: differences in typing, scrolling, and mouse behavior separate AI browsing agents from humans and one another, showing that they are a critical component to reliably detect and control this emerging form of web traffic.