Abstract:

The increased availability and access to generative AI, such as ChatGPT, which can produce high-quality texts, poses significant challenges to educational models that rely on student-generated text for assessment purposes. An expanding body of research addresses issues related to academic dishonesty, alternative assessment models, and the impact on student learning. Concurrently, there is a growing interest in the educational potential of these technologies, where AI can engage students through coaching, providing advice, and analyzing mistakes to support learning. As we have previously reported (Winerö & Lundin, 2024), educators face the ongoing challenge of balancing these potential risks with the possible benefits.

This presentation will provide initial insights from an ongoing Swedish project that collects data from active teachers, focusing on how they navigate this balancing act in their everyday teaching practices.

The research question guiding this study is: What strategies do professional teachers develop to balance the possibilities and challenges associated with generative AI in their work?

The project involves a large cohort of teachers from 13 school units in Sweden, including different ages, and school forms. To be able to engage with a large number of teachers, while focusing on detailed accounts and a development oriented research approach the project will includes both primary and secondary participants. By training primary participants in each unit to conduct local data collection, we aim to gather qualitative data from a larger number of secondary participants. The secondary participants include teachers, principals, special education teachers, educational developers, and others, allowing for a comprehensive understanding of changes in educational activities. We will also discuss the benefits and challenges of this data collection setup.

We approach the teachers’ challenge of balancing risks and possibilities from a systemic perspective. This approach is effective in highlighting how current educational activities are potentially challenged or transformed by the introduction of new AI functionalities and provides tools for identifying opportunities for the development of these activities. Consequently the initial efforts in the project is part of a larger model for transformation, allowing practitioners increased agency in the development of their own practice, as well as the use and design of technology (Fischer, Lindberg & Lundin; 2020). In multidisciplinary research, including critical and design oriented perspectives, new educational futures could be imagined, tested, and responsibly be redesigned. Decisions around such change needs new models, methods as well as an increased professional discussion, involving practitioners to be able to gain understanding of potential futures, but also allowed agency in change processes (Lundin, et al., 2024).

Keywords:

AI, activity theory, design trade-offs, assessment, generative AI, ChatGPT.