Abstract
BACKGROUND: Generative AI (GenAI) such as ChatGPT can take over tasks that previously could only be done by humans. Although GenAI provides many educational opportunities, it also poses risks such as invalid assessments and irrelevant learning outcomes.This article presents a broadly applicable method to (1) determine current assessment validity, (2) assess which learning outcomes are impacted by student GenAI use and (3) decide whether to alter assessment formats and/or learning outcomes. This is exemplified by the case-study on our medical informatics curriculum.We developed a five-step method to evaluate and address the impact of GenAI. In a collaborative manner, the courses in a curriculum are analysed on their assessment plans and together with the teachers, the courses are adapted to address the impact of GenAI usage. RESULTS: 57% of assessments, especially in writing and programming, were at risk of reduced validity and relevance. GenAI impact on assessment validity was closer related to the content and structure of assessments than their complexity according to Bloom's taxonomy. During educational retreats, lecturers discussed the relevance of impacted learning outcomes and whether students should be able to achieve them with or without GenAI.Furthermore, the results led to a plan to increase GenAI literacy and use over the years of study. Subsequently the coordinators were asked to either adjust either their assessments formats to preclude GenAI use, or to alter the learning outcomes and include GenAI use and literacy. For 64% of the impacted assessments the assessment format was adapted and for 36% the learning outcomes were adapted. CONCLUSION: The majority of assessments in our curriculum were at risk of reduced assessment validity and relevance of learning outcomes, leading us to adapt either the assessments or learning outcomes. This method and case-study offer a potential blueprint for educational institutions facing similar challenges.