A Cognitive Usability Engineering Approach to Understanding Bias in Human-AI Interaction
Andre William Kushniruk, Elizabeth Marie Borycki · 2026
Cognitive biases refer to systematic, automatic, and often unconscious patterns of thought that have the potential to adversely affect reasoning and decision-making processes of humans. The application of artificial intelligence (AI) in healthcare has the potential to reduce errors due to human bias. However, care must be taken to ensure that the use of AI does not inadvertently introduce bias that will lead to error. To understand the complex relationship and interplay between humans and AI we propose an approach based on application of human factors and usability research. The approach builds on work in cognitive usability engineering and is grounded in a theoretical perspective from the study of distributed cognition. It is argued that biases need to be understood and mitigated in the context of human-computer interaction. It is further argued that incorporating human factors knowledge in the design, development, and deployment of AI is needed to improve AI accuracy, usability, and useful integration into complex domains.