Are CAPTCHAs Still Bot-hard? Generalized Visual CAPTCHA Solving with Agentic Vision Language Model

Xiwen Teoh, National University of Singapore, Yun Lin, Siqi Li, Ruofan Liu, Avi Sollomoni, Yaniv Harel, Jin Song Dong

34th USENIX Security Symposium · Day 1

This paper introduces **Halligan**, the first generalized visual CAPTCHA solver built upon state-of-the-art **vision language models (VLMs)**. Authored by a team of researchers from Shanghai Jiao Tong University, National University of Singapore, and Tel Aviv University, Halligan fundamentally challenges the long-held assumption that visual CAPTCHAs are "bot-hard" but "human-friendly." By demonstrating the ability to effectively solve a diverse array of unseen visual CAPTCHA challenges without prior adaptation, Halligan marks a significant paradigm shift in the cat-and-mouse game between CAPTCHA designers and attackers.

AI review

This is solid, important research that definitively demonstrates visual CAPTCHAs are now trivially breakable by VLM-powered agents. The 60.7% benchmark rate and 70.6% field study rate aren't just academic curiosities—they're the death knell for visual challenge-based bot detection. The work is rigorous, the threat model is realistic, and the implications are concrete.