From Generation to Matching: The Future of AI in Medical Education

This article was exclusively written for The European Sting by Mr. Isaac Ng, a year 3 medical student at King’s College London. He is affiliated with the International Federation of Medical Students Associations (IFMSA), cordial partner of The Sting. The opinions expressed in this piece belong strictly to the writer and do not necessarily reflect IFMSA’s view on the topic, nor The European Sting’s one.

During my preclinical years, I remember struggling to grasp the management steps for asthma. Frustrated after scouring textbooks with little success, I turned to an AI-powered chatbot for help. It quickly produced what seemed like a comprehensive explanation, but I soon noticed it had conflated long-term and short-term strategies, offering a confusing mix of both, without distinguishing when each was appropriate. What was supposed to clarify things only deepened my uncertainty.

My experience shows a growing concern: while AI can generate vast amounts of content, it often falls short of delivering reliable, accurate, and contextually appropriate information. As we confront these challenges, it’s becoming clear that the future may lie less in generating new content, but more in intelligently matching existing, verified resources to the specific needs of learners. This shift from content generation to content matching could fundamentally reshape how we approach medical education.

The Limitations of Content Generation

AI’s ability to produce content at scale is impressive, but it comes with significant caveats. One of the primary issues is “hallucinations,” where AI generates factually nonsensical content. A chatbot might confidently provide an answer about a rare condition, but the response could still be extrapolated based on misinterpretations or even incorrect correlations. This kind of error is particularly insidious because, on the surface, the information appears legitimate. There is no immediate way to disprove the information unless you know the correct answer, defeating the original purpose. Furthermore, the oversaturation of content not only dilutes the quality of education but also increases the cognitive load on students, making it harder for us to focus on what truly matters.

The Case for Content Matching

Rather than focusing on generating new content, content matching aligns existing, high-quality resources with the learner’s specific needs. This method uses advanced techniques such as semantic search and RAG databases to sift through massive datasets and pinpoint the most relevant and accurate information.

In medical education, where every detail must be precise, this approach guarantees students receive the right information tailored to their current learning objectives. For example, an AI system could match a student struggling with pharmacology with verified resources that specifically address the drug interactions they find challenging, contextualised within the correct medical guidelines for their region.

The Human Touch

Transitioning to content matching involves a reimagining of educational strategies. Educators must move from being passive recipients of AI-generated content to active curators. Employing techniques like topic modelling and entity recognition will truly match the appropriate content for students. This guarantees that the matched content is both accurate and pedagogically sound.

Ultimately, AI should be designed to complement, not replace, the expertise and intuition of educators. The goal is not to create a world where AI and educators work together to provide the best learning experience. Until then, I’ll just keep relying on my year group chat for those tricky medical concepts – because sometimes, the best answers still come from a very clever friend. 

About the author

Isaac Ng is a year 3 medical student at King’s College London. He is currently spearheading qVault.ai, an AI-powered platform for medical students. This initiative earned the Student Partnership Impact Award in 2024 for its exemplary collaboration and impact on student engagement and well-being. Isaac is passionate about integrating AI-generated educational content and exploring innovative teaching methods to address the inequity in education.