Oedipus: LLM-enchanced Reasoning CAPTCHA Solver
Abstract
CAPTCHAs have long served as a defense mechanism to distinguish humans from bots. However, the emergence of Large Language Models with advanced reasoning capabilities poses new challenges to CAPTCHA security. This work presents Oedipus, an LLM-enhanced framework for solving reasoning-based CAPTCHAs, demonstrating the vulnerability of current CAPTCHA systems to AI-powered attacks.
Type
Publication
ACM Conference on Computer and Communications Security (CCS ‘25)
Oedipus demonstrates the vulnerability of reasoning-based CAPTCHAs to LLM-powered attacks, highlighting the need for more robust human verification mechanisms in the age of advanced AI.
