Elsevier Medical Education presents an interview with Marcus Drake, Professor of Urology at Imperial College, on clinical readiness from assessment to practice, and the opportunities of AI in education.
You have great experience with the UK medical licensing assessment and have contributed to Elsevier publications, including Medicine in a Day. From this perspective, what are the main challenges educators face in ensuring that students are fully prepared to become competent doctors in the future?
The Medical Licensing Assessment is superb because it really hones in on very fundamental information. And it’s appropriate in ensuring that the future doctor will remember conditions, because that’s actually how people end up treated. But then there’s also the presentations, which is how people arrive, and the skill of the future doctor is to translate the presentation that the patient turns up with a condition which can then be treated and ensure that the diagnosis is reached efficiently and effectively to give the patient the best chance of a good outcome. So, the MLA is good in providing precisely that.
However, there are many more aspects that a patient could experience by way of both conditions and presentations. There’s also a fundamental environment in which doctors work where governance and information handling and multiple priorities, like continuing professional development, all compete with the thought processes that the doctor has. So, the MLA needs that additional supplement of: how will you truly apply this in the more complex working life that the doctor performs in? There is a lot of additional information which isn’t so readily available to the doctor. Many doctors feel that they end up learning on the job aspects that they feel a bit aggrieved wasn’t properly covered when they were at medical school.
"AI is a great opportunity, and it is absolutely about ensuring that information is accurate, that the best information is made easily available to the learner, and that it is presented in the most effective way."
What role can technology play in supporting faculty? More specifically, how do you see AI contributing to this support?
AI is a great opportunity, and it is absolutely about ensuring that information is accurate, that the best information is made easily available to the learner, and that it is presented in the most effective way. It brings so many opportunities to improve illustrations and to clarify text. Furthermore, for the large number of UK medical students and doctors who are from a diverse range of backgrounds, including linguistic, this could really be a big advantage to ensure the clarity of what is written, which will start out in English, and there’s the opportunity to change to alternative languages at the request of the individual, again, using AI. So, it’s a great opportunity for just getting the best information in a form that will be most effective for the individual that’s doing the learning.
What should the role of the educator be when adopting AI to support student learning?
The educator needs to ensure that it is the learner that’s at the forefront of the learning and that it is the learner that actually benefits and then leads the AI such that the AI is the servant, not the master in the AI-learner relationship. It does worry me a bit that people could be a little bit lazy, rely on AI for quick notes, and not really assimilate the information to a level of depth that ensures that they really are processing it directly in their own cognition. So, the educator must make sure that the learner appreciates that depth of understanding and knowledge needs to be within and not just lip service, in effect, trotting out information that AI has thrown at them.
Considering that more than 75% of medical students use AI tools (Elsevier Research, 2024), what would be necessary to ensure that AI is used with reliable, trustworthy content in a safe learning environment?
There does really need to be a strong validation of information. It’s quite unimpressive how often artificial intelligence ends up hallucinating some ‘fact’ and portraying it in a way that’s suggested as being authoritative, whereas in fact, it’s complete fabrication. I think that’s becoming less commonplace as AI improves and starts to integrate internal checks and validations, but it remains a big risk that there is factual error conveyed by AI. I think a strong emphasis on validation, cross-checking, and elimination of false information has to be a priority. Regrettably, AI can, of course, call on any information that’s published on the internet, whether it’s true or not. So, if it was possible to flag off information as false, that would be very valuable, but also very unrealistic. So, I do fear that AI will have that challenge of presenting false information. That is going to be a possibility for the foreseeable future.
"… the educator must make sure that the learner appreciates that depth of understanding and knowledge needs to be within and not just lip service, in effect, trotting out information that AI has thrown at them."
What is your overall view on the use of AI in medical education today?
I found AI particularly helpful for finding appropriate illustrations. I’ve been very impressed, in fact, how once the AI has appreciated the sort of image that you’re looking for, it will quickly develop or offer several options that it has retrieved from various sources. And that’s been incredibly beneficial. The sheer speed with which it can work has been absolutely fantastic.
Furthermore, it’s also quite nice for unlocking a block. If you have a moment where you’re really just struggling to identify a fact or follow a chain of argument, you can get a suggestion from AI on many occasions that really does unlock and there you are – back to effective activity. So, it’s incredibly useful for productivity, effectiveness, and reliability.
A final, more personal question: how do you use AI in your day-to-day professional work, if at all?
In writing medical textbooks covering a range of clinical circumstances, you do really want to have a coherent and effective group of illustrations to support the information to make it more manageable for the students so they’re not simply reading – they’ve got some visual element that just eeks out the hard work. I’ve found AI has been spectacularly helpful in recent times for getting the options of simpler illustrations to present to students. That, actually, has been the big difference that I’ve noticed recently.
Leave a Reply