Dr Heidi Chumley

From Fire Hose to Automated Sprinkler System 

Heidi S Chumley, MD, MBA

When I went through medical school, we often said learning was like drinking from a fire hose. I still hear medical students say that today. The phrase captures the reality: an overwhelming torrent of information delivered at high pressure, with little time to absorb it all. The sheer volume of material required to become a physician is staggering, and the pace is relentless. 

But what if learning was more like a smart automated sprinkler system? Imagine a system that delivers just the right amount of water at the right time, to ensure optimal growth. That is the vision we are pursuing: precision learning powered by artificial intelligence. 

Internally, we recently partnered with our Innovation and AI team to chart this journey. Our top priority is clear: achieving nearly universal student success through precision learning. This goal is deeply personal to me. Throughout my career, I have worked to expand access to quality medical education for aspiring physicians who were not accepted into the U.S. medical education system. These students are talented and determined, and they make up a significant portion of the U.S. physician pipeline, despite not being admitted to U.S. medical schools. 

But not all students who go to medical school are successful, and "nearly universal success" has often felt out of reach. For the first time in my career, I believe it is possible.   

The Promise of AI in Medical Education 

Think of AI tools as enabling a 100% dedicated academic coach and tutor that holds the combined subject matter expertise of every faculty member. This coach has 24/7 availability. It dynamically creates learning plans, curates content, adjusts modalities, and monitors progress in real time. It can identify when a student struggles and intervene immediately, tailoring the approach to that individual’s needs. 

We are closer than many realize. AI can already analyze performance data, recommend personalized study plans, and retrieve and summarize content. The potential impact on student success is enormous. To unlock that potential, we must overcome several challenges.

The Promise of AI in Medical Education 

Think of AI tools as enabling a 100% dedicated academic coach and tutor that holds the combined subject matter expertise of every faculty member. This coach has 24/7 availability. It dynamically creates learning plans, curates content, adjusts modalities, and monitors progress in real time. It can identify when a student struggles and intervene immediately, tailoring the approach to that individual’s needs. 

We are closer than many realize. AI can already analyze performance data, recommend personalized study plans, and retrieve and summarize content. The potential impact on student success is enormous. To unlock that potential, we must overcome several challenges. 

Challenge 1: Data Quality and Digital Infrastructure 

AI is only as good as the data you give it. Medical school, by design, is not an online program. Instruction happens live. Learning management systems exist but are underutilized because the curriculum was never intended to be digital. 

We need to capture rich, structured data about how students interact with the curriculum. Every lecture, every quiz, every case discussion must generate accurate, capturable, and extensive digital information. Without this foundation, AI cannot deliver precision learning. 

Challenge 2: Balancing Access and Control  
One of the most complex decisions we face is determining appropriate boundaries for AI in medical education. Our guiding principle: AI should augment faculty expertise, not replace it. This means designing systems where faculty-developed content takes precedence, where information sources are transparent, and where human oversight remains essential. In medical education, there are no shortcuts to verification. 

Challenge 3: Building Memory and Continuity 

Finally, effective AI support requires continuity. When a student interacts with AI tools across different courses or semesters, those experiences shouldn't exist in isolation. With appropriate consent and governance, we can work toward more connected experiences. Insights from one interaction can inform the next, helping students build on their progress rather than starting over each time. Done well, this could mean more personalized guidance, earlier identification of students who need additional support, and a learning experience that genuinely adapts to each individual's journey. 

Our roadmap: 

• Digitize the Learning Experience 
  Increase the use of learning management systems and digital assessments. Capture granular data on student engagement and performance. 
• Develop AI-Ingestible Content 
  Structure curriculum materials so they can be indexed, tagged, and analyzed by AI systems. This includes lectures, case studies, and assessments. 
• Pilot Precision Learning Tools 
  Start small. Use AI to personalize study plans for a subset of students. Measure outcomes and iterate. 
• Train Faculty and Students 
  AI literacy is critical. Faculty must understand how to integrate AI into teaching, and students must learn how to use AI responsibly to enhance their own effort. 
• Ensure Ethical and Equitable Use 
  Establish clear guidelines for data privacy, academic integrity, and fairness. AI should be a tool for empowerment, not a source of bias or dependency. 

The Double Helix Approach 

This initiative is part of what I call the “double helix” approach to preparing the next generation of physicians. The strand in this series focuses on teaching students to leverage AI for learning. The other strand addresses how AI will transform clinical practice. Together, these strands form the backbone of a future-ready medical education. 

Closing Thoughts 

The fire hose metaphor has served us well, but it is time to retire it. Medical education should not be about survival; it should be about growth. By embracing AI, we can transform the learning experience from overwhelming to empowering. We can create an automated sprinkler system—smart, adaptive, and precise—that nurtures every student to reach their full potential. 

For the first time, “nearly universal success” is not just a dream. It is within our grasp. And that changes everything.