Generative AI (GenAI) is rapidly entering the clinical space, promising faster insights and operational efficiencies. But for medical professionals, speed alone isn’t enough. Trust, transparency, and clinical relevance are essential.
At Wolters Kluwer, we believe that AI tools must be built on curated, evidence-derived content—not just data—so clinicians can make informed decisions with confidence. As we evolve UpToDate® into a platform that leverages GenAI to support decision-making at the point of care, our focus remains on delivering content that reflects the integrity of clinical knowledge, not just the capabilities of technology.
Provenance is the cornerstone of trustworthy GenAI
Clinicians today face an overwhelming volume of medical information, with new studies, guidelines, and insights constantly emerging. The challenge isn’t just access—it’s discernment.
At the point of care, the allure of fast answers can be tempting, especially in high-pressure environments. But speed without substance is a risk. Clinical decisions—whether verifying the correct dose of a drug or determining the most effective approach to treatment—require precision, context, and accountability. These decisions must be grounded in fact and evidence, not probabilistic technology.
In clinical AI, that means being able to trace a recommendation back to the evidence and expert judgment that informed it. Verification isn’t optional—it’s foundational. That’s why provenance isn’t a technical detail; it’s a clinical imperative.
The three pillars of provenance
1. Origins: Building on trusted knowledge foundations
For GenAI to be clinically useful, it must be grounded in content that is purpose-built for the point of care. That means current, curated, clinically vetted knowledge—not large-scale ingestion of undifferentiated medical literature.
There’s a common misperception that if AI could summarize all medical literature in real time, clinicians would make better decisions. But clinical reasoning isn’t just about synthesis—it’s about knowing which information matters, when, and why. No technology today can make those distinctions independently. That’s why the content that any generative clinical decision support system is built on is critical.
UpToDate Expert AI is grounded solely in UpToDate’s clinical content. The underlying editorial process is not just rigorous—it’s intentional and anticipatory. It’s designed to reflect the clinical scenarios that users face and guide them through with clarity. Every line of content is expert-authored, peer-reviewed, and regularly updated to prevent outdated information from serving as input. Each study added to UpToDate is vetted for its relevance to patient care, alignment with clinical context, methodological soundness, readiness to inform or change practice, and balanced consideration of risks and benefits.
2. Derivation: How insights are transparently created
In clinical contexts, the value of GenAI lies not just in what it produces, but in how that output is derived. A trustworthy system must reflect embedded clinical logic, transparent processes, and source traceability.
Derivation begins with intentional design. Rather than applying generic technology to large volumes of content, systems built for patient care should ideally reflect how clinicians think, prioritize, and make decisions in real-world care settings.
Transparency is equally critical. Clinicians must be able to see how the system arrived at an answer—not just the output, but the reasoning behind it. This includes surfacing the type of logic used, the nature of the question, and the confidence in the response.
And traceability must be frictionless. In UpToDate Expert AI, every generative response includes a direct link to its source—UpToDate—within the same interface. This allows synthesis to improve usability without obscuring the evidence. Verification remains immediate, intuitive, and essential.
3. Accountability: Clinical expertise behind every insight
Accountability starts with ownership. The design of a generative system is not neutral—it reflects the judgment, experience, and intent of the people who build it. Strong systems aren’t just technology applied to content; they’re shaped by teams who understand both clinical practice and the capabilities of AI.
This type of work requires hybrid skills and leadership. It demands fluency in evidence-based medicine, critical thinking, and openness to create meaningful knowledge in new ways. Our approach is grounded in close collaboration between clinicians and technologists, working together. Every element is mapped to the clinical reasoning process and reinforced by safeguards focused on patient safety.
What sets our system apart is the depth and breadth of clinical expertise behind it. At the heart of UpToDate Expert AI is a dedicated team of practicing physicians and pharmacists with backgrounds in specialty care, hospital administration, patient safety, medical education, publishing, and content technology. Their work draws on a broad foundation, the global UpToDate network of over 7,600 contributors—established leaders in their fields—supported by an internal faculty of physician editors trained in evidence-based methodology.
This collective experience is not just a credential—it’s a safeguard. In medicine, having your name on the work signals responsibility, ownership, and a commitment to quality. Conversely, systems that function anonymously risk eroding trust. By making the expertise behind the system visible, we reinforce the principle that GenAI in medicine must be built not only with technical skill, but with integrity.
Our whitepaper, “Building the bridge—Generative AI and the future of clinical knowledge,” covers our perspective on how practical, purpose-built tools like UpToDate Expert AI are building the bridge between shifts in clinical knowledge gathering and innovation. Download it for more insights.
