Xie V, Davis N, Gao A. Auditing Dark-Pattern Risk in LLM Interaction Prompts: Leakage-Controlled Routing, Confidence-Aware Review, and an Autonomy-Centered Audit Design on DarkBench. J.I.T.I.E. 2026;2(4):29-45. Accessed August 23, 2026. https://journal.jci.co.id/jitie/article/view/600