How Human-AI Interactions Shape Collaborative Learning: An Activity Theory Perspective
Wenting Sun, Humboldt-Universität zu Berlin (Germany)
Jiangyue Liu, Suzhou University (China)
Abstract
The increasing proliferation of Generative Artificial Intelligence (GenAI) in society has the potential to reshape classroom ecology. Collaboration with GenAI enriches learners' AI knowledge and skills. However, it remains unclear whether GenAI-assisted tasks facilitate human-human collaborative learning.This study examines the interaction between two activity systems in GenAI-assisted assessment tasks—individual assessment and collaborative assessment—through the lens of the third-generation activity theory[1].This theory provides valuable insights to better understanding the complex interplay among students and GenAI in individual-machine collaboration tasks[2,3].Participants were 36 undergraduate students in a course named “Information Technology and Education”.This experiment was divided into two parts:GenAI-assisted lesson plan assessment by text and collaborative lesson plan assessment by group dialogue.Participants were randomly assigned two-person groups to finish GenAI-assisted individual assessment and then collaborative assessment tasks. The study collected and analysed screen recordings of human-GenAI interactions and speech recordings of the collaborative learning activities. Thematic analysis would be conducted[4].The human-GenAI conversation history and students’ responses to generated outputs from GenAI are thematically analysed to identify human-GenAI interaction profiles. Additionally, themes regarding the propose for which students’ prompt Gen-AI tools during lesson plan assessment in both individual task and collaborative task are identified. Differences between individual and collaborative tasks are compared. The study identifies fluctuations in tools and division of labour elements from GenAI-assisted individual learning to collaborative learning.Understanding the evolution of human-GenAI interaction from individual tasks to collaborative tasks can enable educators to tailor instructional strategies for optimal learning with GenAI as an assistant. This study contributes to the knowledge of integrating GenAI in education, exploring how GenAI can influence collaborative learning, and calls for considerations regarding the construction of GenAI-assisted classroom ecology.
Keywords: Human-GenAI interaction, Activity Theory, Collaborative Knowledge Construction,Classroom Ecology
REFERENCES
[1] Engeström Y., “From mediated actions to heterogenous coalitions: four generations of activity-theoretical studies of work and learning”, Mind, Culture, and Activity, 2021, 28(1), 4-23.
[2] Guo K., Li Y., Li Y., Chu S. K. W., “Understanding EFL students’ chatbot-assisted argumentative writing: An activity theory perspective”, Education and Information Technologies, 2024, 29(1), 1-20.
[3] Woo D. J., Guo K., Susanto H., “Exploring EFL students’ prompt engineering in human–AI story writing: an activity theory perspective”, Interactive Learning Environments, 2024, 1-20.
[4] Braun V., Clarke V., “Using thematic analysis in psychology”, Qualitative research in psychology, 2006, 3(2), 77-101.