Use of Generative Artificial Intelligence and Large Language Models in Radiation Oncology Education and Research

Purpose: The rapid evolution of Large Language Models (LLMs) and generative AI is reshaping various sectors, including healthcare. This project aims to create a repository of curated prompts to facilitate usage of LLMs, making learning more efficient, self- directed, and personalized for residents and faculty.

Methodology: Using Chat GPT, we initially focused on the development and testing of a large prompts repository targeting topics related to radiation oncology education and research. This repository integrates various features and how-to guides, from tailored educational resources and clinical scenario simulations to research support via text-to-code functionality and the summarization of scientific manuscripts.

Results: The project is still ongoing, but at this time we have curated more than twenty different prompts that can help with educational and research endeavors. The prompts available and tested include flashcard generation, providing scientific manuscript summaries and providing multiple choice questions to assess learning, text-to-code functionality to help with basic statistical and data wrangling tasks. We additionally are building a library of best practices that can help with troubleshooting prompts

Conclusions: By the end of this project, we hope to have a website hosting the extensive repository of intentionally engineered prompts targeting various topics related to radiation oncology education and research. In hope to pave the way for future endeavors utilizing such widely available tools.

References: