The convergence of Artificial Intelligence (AI) and Human Factors Engineering (HFE) is the latest revolutionary method to maximize human-machine interaction (HMI) in industrial systems, and efficiency, safety, and reliability are continually the highest drivers of performance. The paper discusses how adaptive interfaces facilitated by AI, grounded on human factors design principles, enhance operator performance, reduce cognitive workload, and deliver automation confidence boost. Mixed-method design was employed with quantitative experiments on 60 industrial operators and qualitative interviews on 20 participants. Quantitative analysis was utilized to compare task performance, error rate, cognitive workload (NASA-TLX), and trust in automation of a baseline HMI (Group A) with an AI-HFE optimized HMI (Group B). Outcomes indicated shorter task completion times for operators using the AI-HFE interface to complete the task, error rates were reduced, ratings of workload were decreased, and trust levels were higher (p < .001). Trust and cognitive workload were also found to be predictors of performance on the task by regression analysis and explained 61% of the variance. Complementary qualitative findings highlighted trends of increased trust, reduced cognitive load, alignment with ergonomics, and training needs. Participants also expressed resistance in the form of concerns about skill loss and dependence on AI. Integrated analysis suggested that AI-HFE systems not only impact enhanced objective measures of performance but also influence subjective perceptions of interaction, with implications for user experience design as well as organizational support. This research adds to the emerging area of intelligent ergonomics by introducing an optimization framework for industrial HMI and with practical guidelines for system designers, engineers, and policy makers. Overall, AI application from a human factors perspective is an approach to improve industrial systems to be safer, more robust, and more human-centered.
Abdulgafar et al. (2025) studied this question.