1.2. University guidelines on GenAI use
Following ChatGPT’s availability, interest in the educational implications of GenAI surged. In early 2023, UNESCO released a report on GenAI in higher education (Miao & Holmes, 2023), outlining several guidelines. Soon after, many universities formulated their own guidelines for teachers and students regarding its use (e.g., see sources [1] to [5] in Table 1). It is common in the existing guidelines to introduce a general description of GenAI technology along with cases of tools, usually with an emphasis on ChatGPT. Ideas for learning tasks that illustrate how GenAI offers opportunities for enhanced learning tasks (e.g., opportunities for immediate feedback on written text) are then elaborated on. The guides also alert about risks and ethical implications, highlighting issues related to bias, reliability, and privacy. Interestingly, some guides offer strategies to teachers for adapting assessment tasks to ensure students remain engaged in active learning, even when utilising these tools. This involves encouraging connections to personal experiences and reflections, data collection, and detailed accounts of how GenAI tools have been applied in completing assignments (e.g., University of Tartu guidelines for using AI chatbots for teaching and studies , 2023). ). However, these guidelines typically stop short of offering specific course policies, leaving these decisions to individual educators. Generally, they reference the university’s existing or expanded ethics policies, which build upon traditional plagiarism procedures. However, these procedures are challenging to apply when GenAI tools are used by students to replace their expected learning actions (Hernández-Leo, 2023), as detection of GenAI use is problematic (Farrokhnia et al., 2023). Hence, it may be worthwhile exploring an alternate approach that can line up with established university academic integrity policies yet is flexible in its application to allow the individual educator to adapt it to their course needs and the current technological landscape. Further, a learning agreement-driven approach can better engage student decision making and motivational processes, aligning with self-determination theory – autonomy, competence, relatedness (Deci et al., 2017) – especially when executed in the context of group work.
This article presents a pilot study involving the use of a learning agreement for GenAI use that was implemented into a university course for first-year students in an engineering degree programme during the 2023-24 academic year. The learning agreement approach involved instructors and students agreeing upon ground rules for AI use at the onset of a course via a multi-page online form. The motivation of the work was to explore whether a learning agreement can be an acceptable tool for governing student use of GenAI in university coursework. As such, the research questions being explored are: R1. Would first-year students be accepting of a course learning agreement on GenAI use? R2. Would students adhere to the terms of the learning agreement in their group assignment? R3. After experiencing a course learning agreement, how would students improve on its use? To address these questions, the article contributes the following: (a) A learning agreement for GenAI use in university courses and literature supporting its terms; (b) A description of modifications made to the course design to account for student use of GenAI and the terms of the learning agreement; (c) Results of a pre-post study exploring changes in student awareness and use of GenAI tools, beliefs related to how such tools should be used in university learning, and thoughts on a learning agreement for GenAI use approach; (d) An analysis of student adherence to terms of the learning agreement on a key course assignment; and (e) Findings from an analysis of student suggestions to improve the learning agreement approach adopted.