AAMC October 1, 2024
To the Editor:
Boscardin and colleagues1 describe optimizing generative artificial intelligence (AI) use in several domains of medical education and close with a consideration of its real-world applications. As medical trainees who currently apply ChatGPT to automate documentation and facilitate clinical decision-making, we wish to underscore the importance of prompt engineering as an essential competency in the effective and safe use of ChatGPT and other large language models (LLMs).
Prompt engineering, the primary means of communicating with LLMs, is the process of creating inputs to generate precise and contextually relevant responses. Well-designed prompts can mean the difference between receiving a focused list of potential diagnoses ranked by likelihood versus a generic list of diagnoses that may not even apply to...