The ghost-writer potential of ChatGPT and related AI tools has been widely discussed in many editorials and essays. Yet, general AI models to-date have failed to fully utilize medical language and this still places some limits to applications in medical science and healthcare [2]. Open-domain question answering firstly attracted the interest of the natural language processing (NLP) community and gave rise to a subset of sources that includes large-scale, multi-subject, and multi-choice datasets for medical domain question answering.
