AI Interfaces for Reducing Cognitive Load in Decision-Making: Dynamic Adaptation of Information Presentation

Complex engineering environments produce heterogeneous information that can overload attention and working memory, reducing decision performance. This study frames information presentation as an adaptive variable and proposes a closed-loop AI interface that adjusts granularity, ordering, modality, and uncertainty encoding using context and interaction signals. The framework targets lower cognitive burden without hiding critical evidence and is evaluated via response time, accuracy, and confidence calibration with machine-learning–based policy learning.