Artificial intelligence (AI) in education has raised several ethical questions regarding academic integrity, privacy, and bias in language algorithms (Master, 2023). These concerns have yielded to distrust to the point that several institutions banned chatbots at Campus locations. Still, the discussion should not be about whether we should use them but how fast we can leverage the data-crunching possibilities to address the biggest challenges of medical education. To do so, educators have the task of transforming the generic use of generative AI to a specific task, such as designing an educational case, a grading rubric, or an exam. However, the quality of the designs depends heavily on the clarity of the educational intent and the technological savviness of the user to provide clear instructions for the algorithm to process.

Early experimentation of educators with generative AI, such as ChatGPT, could be an opportunity to refine the technological savviness and the specificity of prompts we supply to make the algorithm produce more effectively. Recently, Rospigliosi (2023) discussed that the challenge is to define what to ask of ChatGPT, as the quality of the prompt makes the quality of the response. In this ideation and creation stage, where early adopters are piloting small-scale interventions and experimenting with the tool’s limits, educators must work collaboratively across continents and among fields to define guidelines and identify effective strategies for large-scale implementations. While doing so, we speed up the developing process of learning resources in medical education and aspire to have more updated content in the curriculum.

Nevertheless, once a recipe to generate a learning resource has been implemented, the quality of the produced content should be assessed by specialists in the field. The definition of that quality level is also a field for discussion but should be closely aligned with the educational intent. In the future, not so far away, we could even have an AI tool to assess the quality of the content that the generative AI created. Then the question remains: What AI prompt would be the best to do what I, as an educator, need to work on?