Future of Human Performance Modeling in the Era of Artificial Intelligence: Toward AI-Augmented, Verifiable, and Agency-Preserving HPM

Human performance modeling (HPM) has historically converted theories of perception, cognition, and action into executable predictions for design. AI-mediated work now adds prompting, verification, tool switching, adaptive autonomy, and human–AI co-adaptation, creating tasks that fixed operator libraries and hand-authored models only partially capture. This paper introduces a framework for AI-Augmented, Verifiable HPM (AV-HPM), which keeps mechanistic HPM as the causal core, uses AI to accelerate model synthesis and calibration, and requires verification, uncertainty, stress testing, and governance before predictions are used in safety- or mission-critical design. We identify gaps in current practice, specify a layered architecture, and translate the framework into exemplar studies and minimum reporting criteria—a concrete agenda for making AI-era HPM scalable, auditable, agency-preserving, and cumulative across laboratories and domains.