Cognitive Workload and Mental Burden in Health Care Professionals Interacting With AI: Systematic Review and...

Area of Science:

• Health Informatics
• Artificial Intelligence in Medicine
• Cognitive Science in Healthcare

Background:

• AI adoption in healthcare is rapidly increasing, with tools like ambient documentation, diagnostic imaging AI, and clinical decision support systems (CDSSs) becoming common.
• The cognitive demands and verification burden on clinicians using these AI systems are understudied.
• Decision-makers lack structured evidence on AI's true impact on clinician workload and burnout.

Purpose of the Study:

• To systematically review evidence on cognitive workload and burnout in healthcare professionals using AI-powered clinical tools.
• To quantify pooled effects of AI on workload and burnout using a conservative inferential framework.
• To assess the certainty of evidence for different AI categories.

Main Methods:

• Systematic review registered in PROSPERO, adhering to PRISMA 2020 and PRISMA-S guidelines.
• Searched major databases (MEDLINE, Embase, Web of Science, Cochrane CENTRAL) from January 2015 to 2026.
• Used validated instruments (NASA-TLX, PFI) for workload and burnout; assessed risk of bias and GRADE certainty; employed meta-analysis with Hartung-Knapp-Sidik-Jonkman adjustment and prediction intervals.

Main Results:

• Included 21 studies with 2885 professionals; AI's cognitive impact varies by application.
• Ambient AI documentation showed significant reductions in temporal demand and effort, and decreased work exhaustion and burnout prevalence (moderate to low certainty).
• Diagnostic imaging AI and CDSSs demonstrated mixed or increased workload; GRADE certainty was very low for these outcomes.

Conclusions:

• Ambient AI documentation is associated with reduced cognitive workload and burnout, particularly in early-adopter cohorts, though evidence is based on few studies with wide confidence intervals.
• Findings support institutional pilots, regulatory human-factors evaluation, and human-centered AI design.
• The net benefit of AI on the healthcare workforce remains an open empirical question.