New AI Class Helps Medical Students Dissect Artificial Intelligence
Summary
- The University of Miami Miller School of Medicine recently added an artificial intelligence (AI) elective to its combined M.D./M.P.H. program.
- The course offers a comprehensive overview of AI’s impact on medicine and medical research.
- “Introduction to AI in Medicine and Public Health” includes explanations of core AI concepts like machine learning and natural language processing.
Artificial intelligence (AI) is a promising technology that is already having a major impact on medicine. Advanced AI algorithms are being used to detect cancer, take notes during patient visits, summarize the latest medical literature and much more.
To help students keep up with this rapidly evolving tool, the University of Miami Miller School of Medicine recently added an AI elective to its combined M.D./M.P.H. program. The course, “Introduction to AI in Medicine and Public Health,” offers a comprehensive look at AI’s potential, as well as its limitations.
“Seeing what’s happening with AI, especially since the onslaught of generative AI, especially large language models, it’s really exploded into people’s consciousnesses,” said Shirin Shafazand, M.D., professor of medicine and director of the M.D./M.P.H. program at the Miller School. “We need to ensure that our students have the appropriate skills to use AI and understand what these tools actually do.”
The AI in Medicine Course
Dr. Shafazand designed the course to give M.D./M.P.H. students and faculty better information to effectively use AI and understand its inner workings. The course introduces students to numerous AI concepts, including:
• Machine learning (which underpins prediction and prognosis models in medicine)
• Deep learning (a type of machine learning)
• Neural networks (AI inspired by the human brain)
• Natural language processing (which allows computers to “read” and understand human language and generate appropriate responses)
• Prompt engineering (which designs AI queries to retrieve the best responses)
The class dove into ethics, cybersecurity, government regulation and much more. Students learned from experts in academia and industry at the forefront of AI innovation. To cap it off, students were asked to complete a project, such as developing their own AI agent — a tool that can perform specific tasks without human intervention.
“The students came up with some really cool AI agents,” said Dr. Shafazand. “Some were very practical and reflected their interests at this stage of their careers. Many focused on patient education and enhancing medical knowledge. It was nice to see how the students were thinking.”
AI and the Future of Medicine
For many of the students, the class was more than gathering knowledge. It was an ethical imperative.
“AI is only going to become further integrated into society and our practice, and I feel like it’s my responsibility, as a future physician, to get ahead of the curve, see the landscape and understand the ethical implications,” said Alex Pedowitz, a member of the Miller School’s Class of 2026. “How is it already being used in clinical practice, and what can I do as an individual provider to advocate for better and responsible patient care using this technology?”
The class provided unique opportunities to see under the AI hood. While they had all heard of these tools, and maybe dabbled in ChatGPT, this deep dive was eye-opening.
“We started out not knowing how the information was stored or how the data was utilized,” said Chase DeLong, also part of the Class of 2026. “By the end of the course, we understood how difficult it is to create these systems, train them and get one to an operational level while creating the necessary safeguards to protect patient information and identity.”
Using Technology on Behalf of the Patient
Protecting patient data is always a top priority, but making a privacy-compliant AI tool is no easy task. AI can seem incredibly smart, but algorithms lack intuition and can make mistakes. For example, AI can have trouble differentiating between telephone numbers and Social Security numbers, a privacy nightmare.
Still, the technology offers many advantages, and it makes sense for medicine and industry to do the heavy lifting to overcome these barriers. AI-powered notetaking apps can automate electronic health record entries, giving physicians more face time with patients.
“The chief medical informatics officer for UHealth, Maritza Suarez, M.D. ‘04, was one of the guest speakers and demonstrated an ambient scribe, which transcribes and summarizes conversations, live in class,” said Haikel Haile of the Miller School Class of 2026. “It was really interesting to see it take an eight-minute conversation and actually make it into a good note.”
In addition to hearing about AI’s inner workings and potential downsides, the students learned better ways to use the technology. Prompt engineering was a particularly a popular session.
“Prompt engineering can be a really helpful tool to automate tasks and get more streamlined results,” said Pedowitz. “That’s something I’ll definitely use in the immediate future to practice for residency interviews and articulate my thoughts. Later on, as my classmates have already demonstrated with their projects, it can be really useful for patient care.”
The students came out of the class feeling both empowered and forewarned. They know that AI is the future, but it should also be treated with extreme caution.
“We cannot let our human judgment be replaced by AI,” said Dr. Shafazand. “Critical appraisal must always be in play. I want that to be a key takeaway for our M.D./M.P.H. students. Ask the questions and look at AI implementation from all angles. How will it impact society, patients, the health care system and costs? Develop a framework for evaluation and change management and then implement that AI tool into the health care system.”
Tags: AI, artificial intelligence, Department of Medical Education, Dr. Shirin Shafazand, M.D./M.P.H., medical education, medical students, technology