Generative AI such as ChatGPT introduces an innovative approach for assessing answers in radiology examinations, a methodology applicable to candidates in the EDiR exams. Underlying large language models (LLM) show a remarkable quality in the automatic understanding of text that – given one can mitigate risks in bias, hallucinations and misuse – promise productivity boosts for students, teachers and organisations. The European Board of Radiology (EBR) is embarking on an innovative project to enhance the evaluation process of the European Diploma in Radiology (EDiR) by integrating cutting-edge technology. In collaboration with leading experts in the field, the EBR is extending its previous experiments on natural language processing (NLP) using generative AI for exam evaluation. This groundbreaking initiative holds the potential to revolutionise the assessment process of the EDiR, offering a more efficient, objective, and accurate means of evaluating candidate performance.
Chairperson's introduction: AI integration in radiology education
Laura Oleaga Zufiria, Barcelona / Spain
Technical view: overview of generative AI and LLM in radiology
Benedikt Kämpgen, Würzburg / Germany
EDiR's view: revolutionising radiology exams
Laura Oleaga Zufiria, Barcelona / Spain
Clinician's view: how can examiners and submitters benefit from generative AI?
Peter Mildenberger, Mainz / Germany