Lung cancer is associated to high mortality rates and has a large impact on the quality of life of patients and families, who therefore need suitable information to deal with the situation. However, information provided by health services is often not adapted to lay users and retains a considerable number of technicalities that impair comprehensibility. Accessibility, a concept receiving increasing attention, not only involves physical access but also understanding relevant information when it comes to the medical setting. One of the main intralingual translation procedures used to adapt specialised text is determinologisation, which comprises strategies like using common-speech synonyms, explanations, examples, etc. Recent artificial intelligence generative models offer a promising tool to produce texts at different specialisation levels. In this study, we evaluated the capacity of ChatGPT for determinologisation of terms extracted from a corpus of patient-oriented lung-cancer texts and compared the results with reliable patient-oriented online sources. ChatGPT produced definitions and context information similar to those of the online sources in a very short amount of time. However, both the choice of suitable input prompts and the post-edition process necessary to produce quality final texts on lung-cancer information still required the supervision of human experts.
