Paper Number
ECIS2026-1302
Paper Type
SP
Abstract
International, non-native English-speaking postgraduates in English-medium universities often face discipline-specific language barriers and unfamiliar academic norms that inflate extraneous cognitive load and erode confidence. We examine when and how ChatGPT, a Generative AI (GenAI) tutor, can mitigate these challenges. Drawing on Cognitive Load Theory (CLT) and Social Cognitive Theory (SCT), we propose a dual-path framework in which ChatGPT’s affordances (rephrasing, chunking, iterative feedback) reduce extraneous load and foster academic self-efficacy and self-regulated learning. We report qualitative multi-site online interviews with 30 international postgraduates in the United Kingdom (UK) and Australia, analysed via reflexive thematic analysis. This research-in-progress will clarify the mechanisms and boundary conditions—particularly digital literacy and prior educational background—under which GenAI acts as a scaffold rather than a shortcut. We anticipate contributions to theory (an integrated CLT–SCT account) and practice (practical guardrails for inclusive AI adoption in higher education).
Recommended Citation
Ong, Chin Eang and Kayas, Oliver G., "From Overload To Agency: Chatgpt, Cognitive Load And International Students" (2026). ECIS 2026 Proceedings. 3.
https://aisel.aisnet.org/ecis2026/comp_mgmt/comp_mgmt/3
From Overload To Agency: Chatgpt, Cognitive Load And International Students
International, non-native English-speaking postgraduates in English-medium universities often face discipline-specific language barriers and unfamiliar academic norms that inflate extraneous cognitive load and erode confidence. We examine when and how ChatGPT, a Generative AI (GenAI) tutor, can mitigate these challenges. Drawing on Cognitive Load Theory (CLT) and Social Cognitive Theory (SCT), we propose a dual-path framework in which ChatGPT’s affordances (rephrasing, chunking, iterative feedback) reduce extraneous load and foster academic self-efficacy and self-regulated learning. We report qualitative multi-site online interviews with 30 international postgraduates in the United Kingdom (UK) and Australia, analysed via reflexive thematic analysis. This research-in-progress will clarify the mechanisms and boundary conditions—particularly digital literacy and prior educational background—under which GenAI acts as a scaffold rather than a shortcut. We anticipate contributions to theory (an integrated CLT–SCT account) and practice (practical guardrails for inclusive AI adoption in higher education).