The AI algorithm consistently triumphs over CAPTCHA defences Introduction Three AI researchers from ETH Zurich, Switzerland, have adapted an AI-driven image- processing model to by pass Google's reCAPTCHAv2 human-verification system. Research Team Andreas Plesner Tobias Vontobel Roger Wattenhofer These researchers modified the YOLO image- processing model to create a new version ca pable of consistently solving Google's CAPTCHA. Their pa per is available on the arXiv pre print server. Background on CAPTCHA For decades, website administrators have employed various methods to prevent bots from accessing their systems. One widely ado pted technique was CAPTCHA, the Com pletely Automated Public Turing to tell Com puters and Humans A part, which could be seamlessly integrated into login processes. Evolution of CAPTCHA In 2007, Google launched its pro prietary CAPTCHA a p plet, now u pdated to the latest version known as reCAPTCHAv2. Like other CAPTCHA systems,...
Artificial Intelligence Research | Quantum Communication | Space Research Missions | Medical Technology Innovations | Future Science NEWS