How frontline clinicians actually feel about AI entering their diagnostic and therapeutic decision-making matters enormously — because adoption attitudes shape whether these tools improve patient outcomes or gather digital dust. Five years of real-world exposure to AI-powered Clinical Decision Support Systems (AI-CDSS) appears to have meaningfully shifted professional sentiment, and understanding the contours of that shift is essential for health systems planning the next wave of implementation.
This 2025 cross-sectional survey of 215 UK and Italian clinicians used a validated instrument structurally continuous with a 2020 baseline, enabling rare temporal comparisons on attitudes toward AI-CDSS across perceived benefits, harms, regulatory adequacy, clinical utility, and ethical concerns. Cluster analysis using Ward's method identified distinct attitudinal subgroups — segmenting clinicians into adoption profiles rather than treating the profession as monolithic. Key themes tracked included not just enthusiasm or skepticism, but the specific regulatory and ethical concerns that have evolved as clinicians have moved from theoretical exposure to hands-on AI experience in live clinical environments.
The temporal design here is genuinely valuable and underused in health informatics research. Most AI-attitude surveys are one-shot snapshots; this study's comparative architecture allows investigators to isolate how real-world experience — rather than media coverage or academic discourse alone — reshapes professional trust. That said, the 215-person sample spanning two countries limits generalizability, and online self-selection bias likely overrepresents tech-engaged clinicians. The UK-Italy pairing also introduces healthcare system heterogeneity that complicates clean cross-national inference. Cluster analysis, while analytically illuminating, can obscure important within-cluster variance. Overall, this is a carefully constructed incremental contribution: it won't reframe the field, but it provides useful longitudinal signal for implementation scientists and health policy planners tracking whether clinical AI is winning over its most critical audience — the practitioners who must use it.