Perception-Centric Explainable AI: Bridging Cognitive Theories and HCI Design for Enhanced User Experience
Sara Alhasan, Reem Alnanih · Procedia Computer Science · 2025
Artificial Intelligence (AI) systems increasingly influence high-stakes decision-making, necessitating explainability tailored to non-technical users. This research investigates integrating cognitive theories, Human-Computer Interaction (HCI) principles, and Explainable AI (XAI) techniques to enhance user perception, a core cognitive function. A perception-focused framework was developed and tested using a medical symptom checker tool. The study involved 20 participants divided into control and experimental groups. The experimental group used an enhanced tool with features like visual hierarchy and consistency, while the control group used a basic version with minimal explainability. Participants completed pre-test, testing, and post-test phases, including surveys and objective tasks measuring recognition and recall. Results showed that 90% of the experimental group achieved ”Good Performance” in recognition tasks compared to 20% in the control group. Subjective feedback indicated significantly higher un-derstandability, with an average rating of 4.8 out of 5 for the enhanced tool versus 3.1 for the control. These findings underscore the effectiveness of combining HCI principles and cognitive theories, such as Gestalt and Feature Integration, in reducing cognitive load and improving clarity. This multidisciplinary approach advances the design of intuitive, user-centered AI systems, fostering trust and accessibility.