Researchers Explore Advantages and Disadvantages of AI in Medical Education | Newsroom

UNC School of Medicine third-year student Joshua Hale led a study on how medical educators use generative AI. The aim of the study is to better understand how teachers can implement AI tools in the curriculum.

---

AI is a rapidly growing technology that is currently influencing the world of medical education. For educators, this emerging technology is demonstrating its ability to bring real value to classrooms. That said, the use of AI has led to the question of whether these intelligent systems can play an important role in medical organizations. Academia, including medical education, is divided in this new field due to ethical concerns arising from misuse, manipulation and privacy in the use of AI. To better understand the current evidence and best practices for implementing AI that enhances medical education and training, one study — led by third-year UNC School of Medicine student Joshua Hale — aims to educate teachers and students on best practices. implementation of these tools, as well as identifying gaps in the literature to inform future research on why AI should be incorporated into the classroom and curriculum.

“We got some really interesting ideas about how natural AI can augment professors and students,” Hale said. “In particular, many articles have suggested that AI’s natural language processing capabilities will make them ideal candidates for future applications as virtual patients, enabling students to practice independently. and accepting their views.”

Study, published in Journal of Medical Education and Curricular Development, reviewed 521 articles published before June 30, 2023, on how AI is developed and how it can be used and implemented in undergraduate medical education (UME). Forty-one articles were extracted and showed that most of them were opinion pieces, case reports, letters to the editor, editors, and commentaries on the use of artificial intelligence. products while seven articles use qualitative and / or quantitative methods.

The articles are further divided into five categories that acknowledge the role that AI can contribute to UME as well as its challenges: non-clinical learning assistant, content developer, patient interaction, clinical decision-making guide, and medical writing. The results also showed that the greatest potential for AI tools is to be used as a virtual patient and clinical decision guide.

Determining whether these areas will provide the most significant benefits is a question that researchers aim to answer as they seek the integration of AI tools into medical education. The study went on to examine how these five areas are proposed to improve pre-medical education based on the current literature. The “Nonclinical learning assistant” was the first time that articles suggested the use of AI as a guide for students to search and find medical information. The Virtual Assistant ChatGPT 3.5 has shown how it can answer free clinical questions more than 43% of the time. The literature has suggested that AI can be used as a learning assistant or personalized guide for non-medical issues.

The “content maker” appeared in articles about ChatGPT’s ability to write multiple-choice questions and provide explanations. Researchers have found that AI can write multiple-choice questions in the same manner as a human in much less time. For the “curriculum,” it was found that AI can generate cases and comparisons and suggest the implementation of guidelines and tools to prevent misuse by students. Four articles provided insight into how AI models can shape patient-care interactions. It will be able to demonstrate how ChatGPT can act as a real patient and provide feedback to students about their medical decisions as well as a clinical reasoning guide.

“Generative AI can simulate patients who are actually answering questions from students with a specific test or learning objective,” Hale said. “While the actual patients are still there, the artificial AI has the ability to translate text and generate personalized responses based on student interactions. For example, students can conduct interviews with real AI patients on computers. of them, they get follow-up questions, and they get independent feedback on their performance. This technology expands the educational resources available to educators,” he said.

Finally, medical writing was often cited as an area of ​​use and concern. AI has demonstrated assistance with creative writing, research manuscripts, and grant writing as well as clinical note-taking support. To help with the concern of abusing the tool in this group, the researchers suggested the implementation of monitoring systems for the use of AI that generates and organizes assignments to meet its limitations, so students must deal with its limitations.

“The AI-generated content that is generated can evade detection by automation, creating challenges to academic integrity,” Hale said. “Tools like ChatGPT are essential for developing search and feedback strategies. Schools are increasingly valuing these capabilities and educating students on the appropriate use and citation of AI that generates them.”

This study suggests that the ideas presented in these articles can provide a starting point for educators who want to innovate in the classroom and equip students for a medical field that increasingly embraces AI tools. The benefits of these tools can lead to individualized insight into cognitive or practical mistakes a student may make and immediate feedback. By incorporating AI and its new capabilities, educators can spend more time focusing on the next generation of medical professionals and innovators. However, Hale stressed the need for more research.

“Further research is needed to demonstrate the direct benefits of integrating AI tools into classroom learning,” Hale said. “However, these tools are already proving to be valuable additions to interested users. My classmates often use ChatGPT as a study partner to create brainstorming aids, study templates, and more. In a research setting, it can serve as a tool to generate ideas for literature research, learning coding or data analysis, and much more. ”

Content link: Brittany Phillips, Communication Specialist, UNC Health | UNC School of Medicine