Generative AI is steadily weaving its way into every corner of healthcare, influencing clinical decision-making, streamlining administrative work, and redefining how future physicians are trained.
Residency education, in particular, is at a pivotal moment as programs explore how to integrate these tools into their learning models. To discuss what this shift means for medical trainees and educators, Becker’s Healthcare spoke with Claudine Lott, MD, Physician Executive for Clinical Solutions Commercial Transformation and Implementation at Elsevier.
Question: How do you see generative AI transforming residency education, and what unique opportunities does it offer to support residents in developing clinical decision-making skills?
Dr. Claudine Lott: Residents are in a unique position. As early career trainees, they need ongoing education. At the same time, they are practicing clinicians taking care of patients, facing similar challenges as their attending peers.
Given this dual role, residents need tools to support both their learning and their clinical practice. Generative AI has the potential to be very helpful in both areas. Tools for clinical decision support and knowledge retrieval, for example, can support a resident’s practice and research needs. On the other hand, tools like ambient dictation can streamline administrative tasks that get in the way of learning and patient care. Residents can even use generative AI tools to study for in-service exams and prepare for their boards.
While many diverse tools exist, residents need to understand how and when to utilize them. When it comes to clinical questions, for example, residents must ensure they are using generative AI tools to augment their clinical reasoning, not replace it.
Q: In your work with health systems and training programs, what challenges do residency directors face when integrating AI into curricula? How can they overcome these barriers and ensure meaningful adoption?
CL: Residency leaders and faculty, as well as residents themselves, are still learning how to incorporate AI tools into residency education. In an Elsevier-sponsored roundtable hosted by the Society of Hospital Medicine, a diverse panel of residency leaders brought up a challenge that gets to the heart of concerns about this topic: With clinical questions, we want residents to rely first on their foundational knowledge. We don’t want them going straight to an AI tool. Although generative AI can be incredibly powerful, it’s not the be-all and end-all.
To address this challenge, it’s important to create partnerships between faculty and residents. Since residents tend to be younger and may be more accustomed to a digital-first approach, they may be more comfortable with this kind of new technology than the faculty. Faculty need to understand how residents are currently using generative AI and deepen their own knowledge so that they can role model the best ways to use these tools in clinical practice. Access to transparent and reliable tools, such as ClinicalKey AI, can help to support this approach.
Q: Trust is critical when using AI in education. With solutions like ClinicalKey AI, how does Elsevier ensure that recommendations remain evidence-based and align with residents’ clinical judgment and training standards?
CL: At Elsevier, we are committed to responsible AI principles. Aligned with these principles, there are three main components that contribute to making a tool like ClinicalKey AI reliable.
First, ClinicalKey AI is built on a retrieval-augmented generation or RAG architecture. Rather than a single large language model that answers queries solely based on pre-trained data, it searches a curated content set for relevant information and surfaces that information for the user.
Second, ClinicalKey AI leverages high-quality content in its responses, sourcing from an expert-selected collection of trusted, high-impact journals, books, and more. ClinicalKey AI draws on evidence-based information that is aligned with current care standards, and makes the sources clear to users through citations and links. This gives users confidence that the tool is delivering credible information from reputable sources.
Third, ClinicalKey AI uses an evaluation framework based on a clinician-in-the-loop approach. Clinical subject matter experts evaluate system query comprehension and responses for parameters such as completeness, correctness, helpfulness and potential for harm.
Overall, these measures help to ensure that ClinicalKey AI’s responses are grounded in clinical evidence and in alignment with residency guidelines and standards.
Q: Medical residents’ learning needs differ widely by specialty and stage of training. How is Elsevier leveraging AI to personalize educational tools so they’re adaptable across diverse clinical disciplines?
CL: The conversational format of generative AI tools like ClinicalKey AI makes them adaptable by allowing the user to tailor their queries to their individual needs. For example, an early-career resident might ask a more general question, and the tool can point them to resources such as expert-authored clinical overviews and textbook chapters for a broad review of the topic. A later-career resident may ask more complex queries about specific patient scenarios and the system can provide personalized responses that pinpoint sources in the literature. In either use case, the response will be based on reputable content, reinforcing the need to refer to evidence-based sources no matter where the learner is in their career. Additionally, ClinicalKey AI’s curated content set is designed to cover a broad range of specialties, so that learners across disciplines can access the information they need.
Q: Looking ahead, where do you see AI making the biggest impact on lifelong medical education — not just residency, but the ongoing professional development of early-career physicians?
CL: Generative AI is here to stay. Elsevier’s 2025 Clinician of the Future report found that the number of clinicians who use generative AI in their work is growing, and they expect that over time clinicians using AI tools will deliver higher quality care than those who don’t.
Overall, clinicians are optimistic that AI tools will help save time, enhance diagnostic speed and accuracy, and improve patient outcomes. Training residents to use these tools responsibly in the context of their clinical education will enable them to practice more successfully and take better care of patients throughout the course of their careers.
