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

No ratings

Presented at USENIX Security 2025 by

Xiwen Teoh

halligan.pages.devwww.usenix.org

Visual CAPTCHAs, such as reCAPTCHA v2, hCaptcha, and GeeTest, are mainstream security mechanisms to deter bots online, based on the assumption that their visual challenges are bot-hard but human-friendly. While many deep-learning based solvers have been designed and trained to solve a specific type of visual challenge in a CAPTCHA, vendors can easily switch to out-of-distribution visual challenge of the same type or even new types of challenge with very low cost. However, the emergence of general-purpose AI models (e.g., ChatGPT) challenges the bot-hard assumption of existing visual challenges, potentially compromising the reliability of visual CAPTCHAs.