Teaching and testing in the era of text-generative AI: exploring the needs of students and teachers

Abstract

Purpose Large language models such as ChatGPT are a challenge to academic principles, calling into question well-established practices, teaching and exam formats. This study aims to explore the adaptation process regarding text-generative artificial intelligence (AI) of students and teachers in higher education and to identify needs for change. Design/methodology/approach The issue is explored in a mixed-methods approach based on Domestication Theory (Silverstone et al., 1992; Silverstone, 1994), incorporating views of both teaching staff and students. Both statistical and content analyses were carried out. Findings The results show that both students and teachers are conflicted about generative AI and its usage. Trepidation and fear stand against a general feeling that AI is an integral part of the future and needs to be embraced. Both groups show marked needs for training and rules and offer a variety of ideas for new exam formats. Originality/value This study provides a unique insight by exploring the attitudes and usage intentions regarding generative AI of two stakeholder groups: students and teachers. Its results can be of significant use to institutions deciding on their strategy regarding AI. It illustrates attitudes and usage intentions as well as needs of both groups. In addition, ideas for new assessment and teaching formats were generated.