Elsevier Medical Education presents an interview with Dr Sameer M Khan, Assistant Professor of Physiology at the College of Medicine, University of Bisha, medical education specialist, and Elsevier author, focused on responsible AI adoption in medical education, covering curricula, faculty training, ethics, and clinical preparation.
From an institutional perspective, how should medical schools begin integrating AI into their curricula in a structured and sustainable way?
AI has has been very effective in transforming the way in which medical education, as well as healthcare, has been taken forward in recent years. In view of this, I would suggest that institutions should first follow the basic aspects. So, as far as any medical education program is concerned, we always start with what is known as the "program learning outcomes", which I feel should be clear and well-communicated. This would really make a difference in this scenario. Also, the mission and vision statement of the institution as such should be modified to accommodate the AI competencies in terms of digital literacy, technological advancements, and ethical governance. The other important thing I would like to mention here is about the AI competencies themselves. Institutions should focus on integrating AI competencies into their program learning outcomes. At the same time, they should include these AI competencies into the specific learning outcomes of a particular course.
As far as a structured curricula is concerned, the focus should always be on the three major areas in medical education. The first and foremost being the teaching and learning aspects, the second is research and the third is clinical practice. Now, regarding the teaching aspects, more importance should be given to integrating the AI tools in terms of introduction, their usage, as well as application in clinical medicine. In terms of research, the student should be able to understand what the basic research methodologies are and how AI can help them in this regard. In terms of clinical practice, much more importance has to be given to hands-on training, to the use of various tools, as well as drafting a very comprehensive clerkship program to accommodate for the training of graduates and in a sense, helping them out with understanding and appreciating the use of AI in healthcare in general.
So, how to make it more sustainable? Sustainability depends on bridging the gap between the technology as well as strengthening the curriculum itself by integrating AI competencies. Here, we need to focus on a number of things. The first thing is the environmental responsibility. We need to be aware of how AI is affecting the environment in terms of carbon footprint and medical e-waste. Secondly, we need to focus on ethical vigor. The use of ethics and governance has to be very well addressed. Thirdly, we need to focus on the financial constraints. And fourthly, importantly, we have to focus on pedagogical soundness, the authenticity of the information, as well as academic integrity. In view of this, we need to establish a quality control program and a quality assurance program to keep a check on these aspects. Finally, at the end of the course, we need to have feedback – either in terms of continuous feedback or curriculum mapping and evaluation. In this way, I feel that the curriculum can be structured and made much more sustainable.
What are the key challenges faculties face when adopting AI in medical education, and how can they effectively overcome them?
As faculty myself, I do appreciate the use of AI technologies in terms of teaching, learning and research. But at the same time, we also have to acknowledge that there are some challenges which need to be addressed. First and foremost is, of course, the faculty knowledge gap and the training itself. The faculty knowledge gap can stem from time constraints, from an overloaded curriculum, or a kind of hesitancy from the faculty to shift from a traditional setting to an innovative curriculum. These things have to be addressed at the earliest in terms of faculty training programs. These programs have to take into consideration two aspects. One, it should be based on serving the skillsets of the faculty and two, it should be structured in such a way that they meet the academic requirements of the faculty.
The second important thing is the curriculum overload as the curriculum in a medical program is huge and it has to be completed in a stipulated amount of time, which puts extra pressure on the faculty. This can be solved by adapting to a competency-based education method, as well as working on a need-to-know basis. So, some of the important aspects have to be figured out and the curriculum has to be distributed in such a way that it serves the undergraduate as well as the postgraduate education separately.
The third important thing is the ethical concerns. Ethical concerns based on data privacy, based on bias, have to be taken into consideration, and the faculty should be encouraged to learn about it. They should be trained in how to actually recognize the bias, as well as in protecting data privacy.
The fourth important thing is AI hallucinations, wherein the information is false. The faculty should be trained in such a way that they should be able to recognize false information. This requires the faculty to be aware of evidence-based medicine, and they should always be able to differentiate between the false and real information. Accountability issues pose a threat in terms of faculty challenges. So, these accountability issues have to be taken care of, and academic integrity should be maintained at all times.
Another important thing I would like to add here is the positive motivation that should be given to faculty. A positive mindset is essential for the faculty to understand the ongoing changes and try to adapt to the situation. For example, learning to unlearn and upskill is very important for the faculty to be able to cope with the various challenges that they might face in terms of teaching and learning, as well as in terms of research methodologies.
"The importance of AI ethics and AI literacy have to be made much stronger in the minds of the students to help them to apply these things in their clinical practice."
How can institutions ensure that AI-enhanced education aligns with clinical practice needs and truly prepares students for real-world healthcare environments?
The foundation for this begins with the program learning outcomes. However, more importance has to be given to building up the curriculum so that it meets not only the graduates, but also the clinical needs on a larger scale. So, the first thing I would like to focus on is the implementation of a collaborative curriculum, wherein interprofessional education as well as collaboration with the other disciplines is at the top of the list.
The second important thing is that more information and more time has to be given to experiential learning in terms of simulation-based learning, which will have a great effect when the graduate goes towards the clinical side. In terms of holistic learning, I personally believe this should be practiced at all levels, wherein the student is given complete information about a particular topic, not only in terms of theoretical knowledge, but practical aspects as well, and it should start from the very early phases of clinical training itself. So, with this in mind, we talk about how these things can be applied in real clinical situations. For instance, if we take the example of diagnostic medicine, the use of simulation labs and simulation training at the undergraduate level will be taken forward in terms of interpretation, and it will really help the graduate in terms of diagnostic medicine as well. As for personalized treatment and predictive analytics, a sound knowledge about the data – especially big data, bias in data, predictive analytics, and all those things – will really help in combating the global health issues, as well as dealing with any kind of epidemic situations in real-world scenarios.
Thirdly, a close focus should also be kept on the use of real-world AI tools. For example, in the field of remote patient monitoring, students have to be exposed to various tools which are being used, like, blood glucose monitoring and blood pressure measurements. All those things have to be focused on at an early stage, and then they will have to be encouraged to use them when they go into the clinical setting. So, in this way, I think AI-enhanced education and clinical needs can be a good amalgamation to develop quality healthcare in general.
In your book ’Fundamentals of AI for Medical Education, Research and Practice‘, you outline a comprehensive framework for AI adoption across education and clinical care. How can academic institutions leverage this framework to design curricular that are future-ready and aligned with healthcare needs.
For the design of the curriculum and a structured framework, the first thing would be to shift to a competency-based education. The second important thing is to focus on key soft skills, which will form a foundation for clinical medicine. In the earlier years or in the preclinical phase, more importance has to be given to fundamental soft skills such as teamwork and collaborative practice, which will help the student to develop a keen sense of collaborative working and interprofessional capabilities. As the student progresses from preclinical to clinical years, much more importance has to be given to clinical decision-making skills, leadership qualities, as well as working in a team environment. These things will lay the foundation for a clinical setting as such.
Other important things are digital integration or digital interaction in medical curricula, encouraging the students to focus on asynchronous learning and developing an interprofessional approach. These are all valuable for developing a structured framework and implementing it to the needs of the program outcomes and graduate competencies.
As I mentioned in my book, the need for a solid and structured curriculum is very, very important and it has to start at the very beginning of the medical education program. The curriculum has to be kept spiral, or scaffolded, so that it gets repeated over and over again. The importance of AI ethics and AI literacy have to be made much stronger in the minds of the students to help them to apply these things in their clinical practice. So, in this way, I think that the structured framework can be made more comprehensive and maybe more realistic as well.
"Clear-cut guidelines should be established and made available to the faculty and to the students as well, so that they understand their limitations and they understand the guidelines clearly while using the AI tools."
What role should academic institutions play in ensuring the ethical, responsible, and unbiased use of AI in medical training?
Academic institutions are at the center of the transformative change that AI has brought about in medical education and healthcare. Although academic institutions do play a vital role in maintaining the medical curriculum in terms of its implementation and its assessment, with the emergence of AI, they also have an extra role to play. This might include a focus on AI policies and guidelines. Clear-cut guidelines should be established and made available to the faculty and to the students as well, so that they understand their limitations and they understand the guidelines clearly while using the AI tools.
There should be equitable access across all areas to AI tools. More importantly, it is essential for academic institutions to encourage a human-in-the-loop system where human intelligence and human decision-making skills have to be upheld at all times along with the AI-related tools itself.
The other thing I would like to suggest here is the formation of ethical committees. So, ethical committees which deal with the guidelines, as well as the effective usage of ethics in terms of research and clinical practice, have to be set up. These committees can be set up at two levels – at the university level and at the college level. At the university level, it will act as an advisory board which will guide the faculty and the students in terms of ethical issues they might face in practice and in research, and how to mitigate these issues as well. At the college level, the main role of the ethics committee is to regulate the usage of AI by the students. It has to make sure that the students use the AI and know the limitations of the usage in terms of bias, AI hallucinations, as well as misinterpretation of information.
So, by establishing these ethical committees, and by establishing the various AI policies and guidelines, and making the faculty and students aware of these things will be very helpful for the institution to establish ethical governance and to keep a check on how to use AI tools effectively.
For institutions at an early stage of adoption, what practical steps or use cases would you recommend as a starting point to implement AI in medical education?
For early adoption of AI into the curriculum, I would suggest dividing it into three steps. First and foremost is to lay the foundation. Foundation begins with the formulation of the program learning outcomes. Along with this, doing a needs assessment to find out the curriculum gaps – what we have and what needs to be done – as well as focusing on the AI competencies, which we tend to achieve at the end of the program itself. Also, focusing on faculty training and giving the faculty thorough knowledge about the AI curriculum.
The second step includes the establishment of a proper AI unit. This AI unit is responsible for drafting the specific learning outcomes of the AI course which is proposed to be included in the curriculum. We also need to understand the various competencies that we expect from the students and work accordingly with that. Along with that, it is essential to introduce foundation principles of AI in terms of AI literacy, ethics, data analysis and practical use of AI.
The third step is focused more towards shifting into the clinical years. Here, the basic need is to give more space for hands-on training for the students, as well as making them aware of the legal responsibilities and the ethical guidelines that are present, and how students should be able to follow these. This also includes making them aware of their responsibilities as a medical practitioner, and developing a mindset where they need to appreciate that AI is just a helping hand and the decision always has to be a collaborative one. Human interference should be there in terms of clinical decision-making and students need to appreciate the limitations of the AI in clinical medicine and practice.
The other thing I would like to highlight here is the importance of evaluation and feedback. The feedback – either on a formative basis or in terms of course reports and program evaluation – should be taken, so that it serves two important things. It serves us to know about the educational gaps that might have occurred in implementation. It also serves as a quality assurance document, so we will be able to know how the curriculum is progressing and what needs to be done in terms of quality.
What core AI-related competencies should faculties prioritize to effectively prepare future healthcare professionals for an AI-enabled clinical environment?
First and foremost is thorough knowledge about AI literacy and the workings of AI technologies, data science, and statistical aspects. All these things have to be thoroughly embedded in the medical graduate. Second of all is thorough knowledge about AI ethics and how these ethics can be used more proficiently. Knowing AIs limitations is very important for a medical graduate to understand and apply. The third important thing would be critical appraisal. The medical graduate should be able to critically appraise the use of AI in clinical medicine. The fourth is clinical applicability and workflow adaptability of the AI tools. The medical graduate should be able to appreciate the use of AI as well as appreciate the workflow adaptability of the AI tools in terms of diagnosis, in terms of changing data sets, etc. The other thing to focus on is the interdisciplinary or interprofessional collaboration. The graduate should be encouraged to engage with interprofessional education, to get adjusted to the interprofessional teamwork, and to understand the team dynamics to perform more effectively in a clinical setting. Last but not least is digital proficiency. Digital proficiency has become one of the most important things to be addressed, and one of the most important things to be taught to a medical graduate. The more recent competency which is being added is prompt engineering. Prompt engineering is needed for the medical graduate in terms of identifying the prompts, identifying the accuracy of the prompts, analyzing the responses from AI tools in relation to the prompts asked, and checking the validity as well as the authenticity of the reply.
Leave a Reply