Abstract
Educators face a growing tension between the efficiency gains of generative artificial intelligence (GenAI) and the preservation of work autonomy. This study proposes Contextualised Shared Agency (CSA), a framework for human–AI interaction, in tertiary education, to uphold educator autonomy across varied tasks. Our CSA model explains how responsibility and control can be dynamically shared between educators and GenAI along two dimensions, i.e., protocol development and action execution for different categories of education tasks. We demonstrate CSA’s utility by conceptualising four GenAI-enabled features, each with a deliberately calibrated agency distribution between human and GenAI, co-designed with and tailored to educators in tertiary education. These illustrate how GenAI can augment educators in task-sensitive ways to protect professional humanity. This work contributes one of the first contextualised frameworks for shared agency in GenAI-based educational technology, aiming to guide the development of AI tools that amplify teacher capabilities while preserving human autonomy.
Recommended Citation
Zhuang, Amy; Shi, Yunfei; and Toorn, Christine Van, "Contextualised Shared Agency: An Action Design Research for
Generative AI-powered Education Tool" (2025). ACIS 2025 Proceedings. 104.
https://aisel.aisnet.org/acis2025/104