The adoption of artificial intelligence in healthcare systems worldwide has accelerated significantly, yet the use of unregulated “shadow AI” tools poses substantial risks within clinical environments. A survey involving over 500 U.S. healthcare professionals revealed that 40% had observed colleagues utilizing unapproved AI applications, while 17% admitted to using them themselves. Yet alongside this progress, a growing challenge has emerged: the proliferation of these tools in clinical environments introduces risks around data security, compliance, and clinical reliability. Among respondents, 23% identified patient safety as their primary concern, highlighting the conflict between AI-driven efficiency and the need for reliable, trustworthy systems.
This challenge is not limited to the United States. Hospitals often rely on multiple external AI tools, creating fragmented adoption that introduces inefficiencies and inconsistencies in patient care. The solution requires seamless integration of governed, evidence-based AI directly into clinical workflows, rather than permitting unchecked tools to spread unchecked. This dynamic is increasingly relevant across the world, including in the Middle East, where digital health initiatives are expanding rapidly. That fragmentation and potential for unwarranted variability in care is only increased by the proliferation of AI tools in clinician workflows.
Wolters Kluwer’s UpToDate Expert AI provides a response to this issue by embedding AI capabilities within existing clinical decision support systems. Unlike general-purpose AI models, it retrieves and synthesizes information from UpToDate, a widely trusted medical reference platform. This ensures that responses are grounded in evidence and clinically validated, while still supporting, not replacing, clinical judgment. The system incorporates transparency in sourcing and governance, addressing key concerns about shadow AI. It provides a faster, conversational way to access expert-authored UpToDate content, offering a trusted alternative to general-purpose AI tools.
Burjeel Hospital in the UAE was among the first institutions to test this technology. Dr. Halprashanth D. Sathyanarayana, a consultant neurologist there, has relied on UpToDate for over a decade. Recently, he used the platform to assist in diagnosing a young patient presenting with acute motor weakness. “I have been a regular user of UpToDate for over 10 years, and it has consistently supported my clinical decision-making, particularly in complex cases,” he explains.
He recalls a challenging case involving a young patient presenting with acute motor weakness: “Using UpToDate, I was able to review key concepts … which provided an important clinical clue.” This led to a critical diagnostic shift, from an initial suspected Guillain-Barré syndrome to a rare metabolic disorder, Multiple Acyl-CoA Dehydrogenase Deficiency (MADD). The patient ultimately received targeted treatment and made a full recovery. Dr. Sathyanarayana notes that this particular case highlights how “UpToDate facilitated timely diagnostic reconsideration and guided appropriate management, ultimately improving patient outcomes.”