Skip to main content
MedEdMentor

Theory-Based AI for Medical Education Curriculum Development

Using AI to substantiate your education ideas

Authors: Gregory Ow
Editors: Geoffrey V. Stetson, MD
Read Time: 7 min

Introduction: The traditional flow of ideas in medical education

In this lesson, we'll explore AI-assisted curriculum development.

We'll start with the historical flow of ideas from medical educators to medical students, and then move on to the transformative potential of the internet and artificial intelligence to completely change what this flow looks like.

Then we'll move onto Kern's Model of Curriculum Development, and we'll give you a prompt that will show you how rapidly AI can generate ideas for you in the context of Kern's Model..

Before diving into these topics, let's understand the classic flow of ideas in curriculum development.

From idea creation to consumption: The traditional five-step sequence

This concept is from media, but applies very well to curriculum development.

In the context of education, particularly medical education, here's how the idea flow traditionally works:

  1. Idea Creation: The educator has the idea of what to teach, often identifying gaps in existing material or seeing the need for a new approach.
  2. Substantiation: The initial idea becomes concrete as the educator develops tangible content such as lesson plans, presentations, or assessments.
  3. Duplication: The tangible content is copied. Historically, this involved making copies of textbooks or educational materials. In today's digital age, it's as simple as copying a file.
  4. Distribution: The content is distributed to learners. Historically, this meant orally delivering the curriculum with lectures or disseminating textbooks and other written material. Now, it often involves digital distribution through e-learning platforms and could consist of text, image, audio, video, 3D models, and others.
  5. Consumption: Finally, the learners interact with the curricula.

From the educator to the learner, this five-step sequence offers a foundational understanding of how ideas flow from creation to consumption.

Evolving bottlenecks in the idea flow

Historical challenges in duplication and distribution

In the early days of humanity, duplication and distribution were the greatest challenges. Knowledge was first transmitted orally, which meant that the only way to share knowledge was to talk to as many people as possible (e.g. think of the Roman Forum).

Then, systems of writing and then printing developed, and printed media was distributed much farther and wider.

However, from an education standpoint, the flow of ideas from educators to learners (and vice-versa) was still severely constrained. Early medical texts were handwritten on papyrus (!!), and then even with the creation of printed media like books and journals, written materials still needed to be physically moved from place to place.

The internet and computers change the game

However, the advent of computers and the internet completely changed this dynamic. Suddenly, creating duplicates of educational content became as simple as copying a file, and distribution was just a click away.

These technological advances significantly reduced the bottlenecks in duplication and distribution. Anyone can have a thought, record it in writing/audio/video, and then send it around the world instantly. In fact, with video conferencing, you could argue that duplication and distribution are happening instantly.

The current bottleneck: Substantiation

In medical education, if you've recorded a lesson for your learners, you no longer have to physically go and talk to each of them, you can just send it to them electronically.

No longer is duplication the hardest part of the idea flow. The most difficult step is turning the raw idea into an actual, consumable curriculum in the first place. That is, the hard part is substantiating your lesson.

This shift of bottlenecks sets the stage for how...

AI can simplify and expedite the substantiation phase

Just as computers and the internet significantly reduced the costs of duplication and delivery, AI will significantly reduce the cost (cognitive energy and time) of substantiation.

Let's now shift to curriculum development, and you'll see what I mean.

Curriculum development and substantiation

In medical education, curricular development (Curriculum Development and Evaluation) often follows a structured approach like Kern's Model of Curriculum Development.

This model is commonly used to take educators from identifying a problem to evaluating the outcomes of a new or revised curriculum.

The traditional way: labor-intensive substantiation

  • Problem identification: Traditionally, educators invest at least some time to pinpoint the gap they want to work on.
  • Targeted needs assessment: After a problem area is identified, some more time is spent evaluating the specific needs of the target audience. This usually involves distributing and analyzing surveys, conducting interviews, or even organizing focus groups.
  • Goals and objectives: Once the needs are clear, educators create learning objectives, typically with Bloom's Taxonomy.
  • Educational strategies: Educators then explore various pedagogical methods, from lectures and seminars to experiential learning.
  • Implementation: Rolling out the curriculum has become much easier recently, as stated above, but there are many organizational and political steps here, such as getting stakeholder buy-in, for example.
  • Evaluation and feedback: Finally, the curriculum undergoes an ongoing evaluation process that involves collecting feedback, conducting assessments, and revising the curriculum based on these insights.

AI-augmented curriculum design: endless substantiation

Here's how AI could speed up curriculum development.

  • AI-assisted problem identification and needs assessment: AI-driven analytics tools can sift through vast amounts of data far more quickly than a human, identifying gaps and opportunities in existing curricula.
  • AI in targeted needs assessment: AI can not only more rapidly analyze the needs assessment data, it could also reduce the need for surveys (external PDF)!
  • Streamlining goals and objectives: Guided educators could use AI tools to easily follow Bloom's Taxonomy to generate learning objectives.
  • AI-optimized educational strategies: We will discuss this below.
  • Implementation: Possible for some AI-related innovation, but usually not too onerous of a step.
  • Automated evaluation and feedback: AI could automatically compose surveys, gather and analyze student performance data, offering real-time insights that can be used for continual improvement of the curriculum.

AI-assisted educational strategies

We're now going to focus specifically on the educational strategies step in the Kern model.

Let's explore a practical example using a ChatGPT / Claude prompt to highlight how AI can not only generate different educational strategies but also adapt quickly based on user feedback.

Introduction to a ChatGPT / Claude prompt focused on educational strategies

Imagine you'd like to improve how to teach hyponatremia to 3rd year medical students.

You've done your needs assessment and set learning objectives, and now you're focused on choosing an educational strategy. Many educators may pick an educational strategy they're most familiar with, or maybe try a new one out.

But what if you'd like to see how 3 different education strategies would look like?

Traditionally you'd either have to compare them in your mind, or spend some time designing each format and then comparing them.

Here's where AI, with your guidance, can substantiate ideas for you. We'll be using 3 theories from our Ten Essential Theories:

Hello, you will act as LessonPlanGPT. Your task is to help me develop my lesson plan. We'll frame the lesson plan through the lens of three different educational theories. **Educational Theory 1: Constructivism** Create a lesson plan that enables learners to actively engage with the topic to build their own understanding. What kind of active learning strategies could you incorporate? **Educational Theory 2: Experiential Learning** Develop a lesson plan focusing on hands-on activities or simulations, emphasizing real-world application. How can the learning experience be made immersive and reflective? **Educational Theory 3: Reflective Practice** Design a lesson plan aimed at fostering self-assessment and reflective thinking among the learners. What tools or activities could you include for this? Instructions: 1. Please silently read and consider each theory. 2. Silently review the below source material. 3. For each theory: --- Think step by step how to generate a concise lesson plan. Make sure each plan is both informative and engaging, and well-suited for my audience. --- Output the lesson plan --- Explain how the education theory is demonstrated in your plan Audience: 3rd-year medical students on an internal medicine rotation. Topic: Hyponatremia Lesson Plan Notes (optional): { I pasted notes on hyponatremia here }

Constructivism

Experiential learning

Reflective practice

Guiding the AI output

Okay, that was quite efficient, taking just 1 minute. You could expand this to cover 5 theories, or maybe even all Ten Essential Theories.

Now, look how AI can assist you further by providing it with new instructions as you pick and choose your favorite parts:

I like the think and pair share from the first one. The simulation one takes a long time, let's keep just the simulated serum and urine Osm measurements. I think the reflective writing piece is good, maybe we can try merging it with the pair share? Can you take the parts I like and come up with a new lesson plan, and then the theoretical basis as well of course.

Integrated lesson plan

Your feedback refines the AI's output, enabling a form of co-creation that can drastically speed up the substantiation phase.

It's an exciting era for medical education; imagine the good ideas waiting to be substantiated!

Limitations of AI in this context

While AI's capabilities are groundbreaking, it's essential to be aware of its limitations. AI can handle basic content creation but lacks the nuanced understanding and clinical expertise that a human educator brings. It is also prone to biases. It's a powerful tool, but one that serves best when used in tandem with human skills and judgment.

The need for theory-based AI

The importance of educational theory in AI-assisted curriculum development

In our example, the use of AI was always grounded within the framework of educational theories. We first anchored our curriculum development approach in Kern's model, and then cycled through multiple educational theories to develop various teaching strategies.

This grounding in educational theory does several crucial things:

  1. Quality: It ensures that the curriculum being developed is not just a haphazard collection of content but is systematically designed to achieve specific learning objectives. This adds an extra layer of quality assurance that would be absent if we just relied on AI to generate content.
  2. Richness of educational approaches: By running the AI through the lenses of different educational theories, we unlock a richness of pedagogical approaches. This step would not automatically have been done without framing the prompt as theory-based.
  3. Rapid iteration: The speed of AI's assistance allows us to rapidly iterate through different theoretical lenses, almost instantaneously substantiating initial ideas into detailed curriculum plans. When combined with the depth of insight provided by grounding these iterations in educational theories, we achieve a powerful synergy for curriculum development.

These aspects collectively argue for a theory-based AI approach in curriculum development. In essence, the theory-based AI acts as a bridge between rapid technological advancements and pedagogically sound educational experiences, offering the best of both worlds.

Conclusion: A new era of opportunities

Traditional curriculum development, while rich in education theory, often lacks the scalability and adaptability that technology can offer. In contrast, AI-powered tools alone can generate content quickly but often lack the nuanced understanding of learner needs that educational theory provides.

So, think of all the innovative courses, creative teaching methods, and impactful lessons that are currently just ideas, waiting to be brought to life. There's no better time than now to turn those ideas into reality.

Take home points

  • The significance of bottlenecks — Understanding the historical bottlenecks in medical education, from duplication to distribution, helps us appreciate the current opportunity with diminished substantiation cost.
  • Quality and depth through theory — Grounding AI in educational theory ensures that the curriculum is systematically designed, adding a layer of quality assurance and theoretical depth that would not be possible through AI alone.
  • Educational diversity — Utilizing multiple educational theories in the AI-driven process allows for the creation of materials that can cater to various learning styles and preferences.
  • Speed and customization — AI's ability to rapidly generate content, when filtered through the lens of educational theories, allows for fast iterations and customization, creating a powerful tool for curriculum development.
  • Synergy for impact — Theory-based AI combines the advantages of technological speed and educational depth, offering a balanced approach that stands to revolutionize curriculum development in medical education.

Related Glossary Pages