Impact of AI-Assisted Radiography on Diagnostic Accuracy and Workload

Abstract

The integration of artificial intelligence (AI) into diagnostic radiography has rapidly transformed clinical imaging practices. AI-assisted radiography refers to the use of machine learning algorithms and deep learning models to support image interpretation, detect abnormalities, and prioritize cases. Proponents argue that AI can enhance diagnostic accuracy while alleviating clinician workload, yet empirical evidence remains mixed, and the mechanisms linking AI use to clinical outcomes are not fully understood. This study investigates the impact of AI-assisted radiography systems on diagnostic accuracy and radiology workload, examining radiologist trust, workflow integration, and perceived ease of use as mediating and moderating factors. A cross-sectional survey was conducted among 720 radiologists and radiologic technologists across tertiary care hospitals adopting AI tools. Five latent constructs were measured: AI Integration Strength, Perceived Ease of Use, Radiologist Trust in AI, Diagnostic Accuracy, and Workload Impact. Data were analyzed using statistical tools structural equation modeling to assess direct, indirect, and interaction effects. Results indicate that stronger AI integration positively predicts perceived ease of use (β = .457, p < .001) and radiologist trust in AI (β = .399, p < .001). Perceived ease of use significantly influences diagnostic accuracy (β = .378, p < .001) and reduces workload impact (β = −.342, p < .001). Radiologist trust moderates the relationship between ease of use and diagnostic accuracy (β = .185, p = .002), suggesting that trust enhances the positive effect of AI tools on diagnostic outcomes. Mediation tests reveal that perceived ease of use partially mediates the relationship between AI integration and both diagnostic accuracy and workload. These findings demonstrate that successful implementation of AI in radiography is not solely a technological endeavor but is shaped by user perceptions and trust. Hospitals should focus on training, usability optimization, and trust-building strategies to maximize AI’s clinical benefits. Enhancing diagnostic accuracy while reducing radiologist workload can improve patient outcomes, throughput, and staff satisfaction. The results provide actionable insights for policymakers, clinical leaders, and informatics teams designing AI deployment strategies in imaging departments