The Obvious Invisible Threat: Vulnerabilities of LLM-Powered GUI Agents to Adversarial UI Manipulations in Web Interaction Tasks

Published in ACM Transactions on AI Security and Privacy (TAISAP 2026), 2026