1179 Navigating the AI Landscape: Surveying the Use of AI Language Tools for Medical Portfolios

Abstract

Recent advances in artificial intelligence (AI) driven natural language processing have made it possible to use tools such as ChatGPT to generate pieces of text which are realistic, eloquent, and challenging to distinguish from human-generated content. This work aims to answer the question: could the use of these tools for medical portfolios represent a positive opportunity to enhance learning, or would it detract from the purpose of reflective practice? In total, 98 post-graduate doctors from across the UK were surveyed to gather data on their awareness of, previous use of, disclosure of use and experience with these tools for their professional portfolios. Their opinions on the need for future guidance in this area were also explored. From 90 respondents who currently maintain a professional portfolio, 18 (20%) had used AI-driven tools in some form to generate content for it and none disclosed this use. Regarding disclosing the use of AI-driven tools, 67.3% advocate for this whereas 32.7% believe disclosure is unnecessary. There was slightly stronger consensus regarding how useful guidance from regulatory bodies on this practice would be, with 73.5% in support of guidance and 26.5% in opposition. Clearly, there is an urgent need for larger studies, discussion, and consensus from regulatory bodies so that guidance can be put in place. Without this, maintaining the integrity of CPD will become challenging for both trainee doctors and assessors of medical portfolios.