Different lenses on human-AI collaboration: towards a comprehensive view on hybrid intelligence in education

This paper aims to deepen our understanding of hybrid intelligence in education by investigating human-AI collaboration in four real-world educational use cases. We argue against the simple view of automation and augmentation as distinct processes and foreground the automation-augmentation paradox. This paradox proposes a comprehensive view of automation and augmentation as co-occurrent processes that drive task allocation between humans and AI. We use this paradox as an analytical lens to examine patterns of task changes in human-AI collaboration across four educational use cases. This study approaches teacher-AI collaboration as emerging from task-level changes when AI tools are implemented in educational settings, and these task-level changes are conceptualised as replacement, complementation, and transformation. Across cases, the analysis suggests that automation and augmentation are distributed across different stages in the instructional process. Replacement was primarily situated in bounded analytical subtasks, whereas complementation emerged when teachers interpret, validate and enact AI-generated output. Transformation was most visible in the planning and reflection phases. The cases illustrate how AI may redistribute tasks between teachers and AI systems without displacing teachers’ pedagogical judgement. These findings suggest that the automation-augmentation paradox can serve as a useful analytical lens for analysing meaningful patterns of human-AI collaboration.