Patterns of Adoption and Learning: Students’ Relationships with Generative AIs

Abstract

Generative AIs have become part of increasingly intimate relationships between humans and technologies, and have created both serious concerns and heightened pedagogical interest in higher education. However, though interest in generative Ais and their contribution to education is spreading, little is still known about how they are used by students in their everyday lives and how this affects education. In this paper we investigate how GAIs such as ChatGPT enter students’ lives and become part of their learning. The paper draws on observations and interviews with students in a Master Program where students partnered with ChatGPT to investigate concepts and philosophical aspects of technology. The aim of the study was to understand how ChatGPT could support students’ collaborative group work as part of problem-based learning practices. As our study involved investigations of students’ everyday uses of generative AIs as well as their learning we were able to make connections between students’ learning strategies and their emerging experiences with GAI technologies. In the paper we investigate these relationships focusing on how GAIs are enrolled and participate in students’ lives and in collaborative learning contexts where uses and understandings of GAI technologies are negotiated in group sessions. Our data suggest that some students enter education with extensive experiences with generative AIs and others commence their engagement after meeting these technologies through education. What characterizes these patterns of adoption - and what are their effects on learning? Theoretically, we draw on sociomaterial approaches to understand GAIs as material agents in students’ lives and in education building on the concepts of patterns of relations and distributed agency. These concepts emphasize the collaborative relationship of humans and GAI technologies, underlining both the specifics of GAIs as ‘human-like’ agents and the blurring of agencies and authorship involved in e.g. AI-generated writing.