AI Uses at OCOM

It has been less than three years since ChatGPT released its large language model (LLM) artificial intelligence (AI) platform. Over this period of time, a lot has changed. The hype for AI was huge at that time, and this technology is actually living up to the hype and improving the lives and work of so many people. 

I see AI being used everywhere. Healthcare is adapting quickly with some obvious use cases (e.g., physician documentation, predictive analytics), and the general public is starting to adopt it for their everyday search and research needs. It seems like every healthcare system now has a Chief of AI physician leader and universities are starting to offer AI programs. 

Medical schools are also rapidly adopting this new technology. Many have an AI committee to advance ideas related to this technology, and some are starting to offer electives for medical students interested in learning more about AI. 

For this blog, I thought it would be interesting to discuss how OCOM is currently using AI in the following ways:

Notebook LM for Creating Podcasts of Educational Material

For folks new to AI, I often recommend that they check out NotebookLM by Google. We are fortunate to use Google Workspace as our productivity suite at OCOM (best in class in terms of collaboration and security) - it has the best AI tools available to an organization on the market. This suite includes NotebookLM. It is a remarkable product - you can upload materials, and it can create study guides, FAQs, briefing documents, and even a podcast of the material. The podcast is super cool. The virtual hosts of the created podcast discuss various facets of the materials that you have uploaded. Our students and faculty use Notebook LM all the time, and it is an incredible example of how modern medical education is adapting at warp speed to the new technology. 

Creating Test Questions 

AI can easily create high-quality test questions from any material or learning objectives in seconds. Our faculty has used it to help the item-writing process, and our students use it all the time to help create study questions from learning objectives and materials. Answering questions is an example of active learning and can help medical students learn materials more efficiently. 

Grading SOAP Notes

Physicians do their documentation in a structured format called a SOAP note (SOAP stands for Subjective, Objective, Assessment, and Plan). Having graded thousands of these over my career, I can verify that grading SOAP notes by medical students is a very tedious and laborious task. Our faculty are using AI to help grade them. While the AI does a good job of assisting, it hasn’t replaced the expertise of the faculty as well as the faculty’s role in making it a formative process for the students. While some physicians are starting to use AI to write their SOAP notes, it is important for a physician to be able to write and review their own SOAP notes to effectively ensure that the ones generated by the AI in the future. 

Summarizing Assessments

Like all medical schools, we have a lot of surveys, feedback forms, and other types of assessments. We have been using Gemini to create detailed summaries of the data, including averaging Likert scales, including quotes, and providing overall summaries of the feedback. I personally use it for this purpose all of the time. For example, we now have a State of OCOM monthly all-hands meeting where I ask the team to provide feedback for the session that I share back with the group. In the past, I would have spent considerable time manually summarizing the assessment and Gemini now does that for me in an objective way in just a few seconds.

Letters of Recommendation

As a medical school dean, I constantly write letters of recommendation (e.g., colleagues/faculty/staff for awards, faculty/students for nominations to national committees, etc.). Sometimes these letters take me up to 30 minutes or more to write. With Gemini, I can input all of the personalized information related to the recommendation that I want in the letter and AI will write an excellent and well-written draft for the letter in just seconds. I have found that it saves me time and improves the quality of the letters as well as focusing my time on the substance of the letter that is the most meaningful. 

Annual Evaluations & HR Issues

I can take my detailed notes and create detailed and thoughtful annual evaluations and corrective action plans on employees in a fraction of the time that it used to take me. I have found the style that HR uses to be careful in its wording around sensitive situations. I can ask it to phrase corrective action plans in a hopeful and optimistic tone that captures the spirit I want to go for. Like all AI assistance, the key is to thoroughly edit everything it creates and ensure that it truly captures the intention and facts. 

Creating Spreadsheets of Data

I am now using AI to help create spreadsheets from various sources of information. For example, I wanted to catalog information on the OCOM website into a spreadsheet recently and I simply asked Gemini to do it for me. In seconds, it had created an excellent Google Sheets that would have taken a human significant time to do and it did it without errors. 

Assistants with Writing Standard Operating Procedures (SOPs), Policies, and FAQs

As a relatively new medical school, we have had to write numerous SOPs, policies, and FAQs. Knowing exactly what you want in these documents and writing very detailed prompts, Gemini has helped us create some of these documents in a fraction of the time. I would also say that AI has, at times, brought in various aspects of policies and SOPs that I would have not thought to include and have arguably improved the quality of the final documents. 

Interview Feedback based upon Rubrics

One way that I hear faculty and staff use AI is to help summarize their interview feedback (employees and student applicants), based upon our established rubrics. They will typically take their notes and input them into Gemini and it will create nice narratives that are objective and based upon the information input. 

Email Composition

I think that my entire team now uses AI to help write emails, particularly those around sensitive topics. We can even ask AI to have a particular tone to the email (e.g. positive, persuasive, etc.). It can also save time writing complex emails. 

Research & Scholarly Activities

We have team members who use AI to assist in the development of research. They use it in various ways: formatting citations, helping with compositions and rewrites, and submission processes. One team member is an expert in an area and used AI to help author a textbook chapter, significantly reducing the time it took for him to do so (and helping overcome a writers’ block). 

The above are just some current examples of how we are just starting to scratch the surface in terms of how AI will impact the future of OCOM and medical education. I think this is the most exciting time ever to be in medical education and AI plays an important role in it. We have created OCOM in such a way to remain a leader in modern medical education in the years ahead. 

Please note that I do not use AI in the writing of any of my blogs. 

Watch ‘The OCOM Podcast’ on AI

Available on YouTube, Spotify, and Apple Podcasts

Click here to watch now.