Strategies for AI Use in Thesis Writing among Undergraduate Language Students: A Think-Aloud Protocol

Abstract

Thesis writing has always been regarded as a complex process. However, the emergence of Artificial Intelligence (AI) is assumed to be helpful as it would enable students to simplify the process of thesis writing. The use of AI tools in thesis writing among students is becoming more prevalent, but there is limited understanding of how these tools are being integrated into the writing process and how students balance their own involvement with AI support during the writing process. Thus, there is a need to explore the strategies students employ when interacting with AI tools, and how these strategies impact their writing process. Understanding these dynamics is essential for improving the way AI can be utilised in thesis writing. This study aims to explore areas of AI utilisation in undergraduate students’ thesis writing and the AI tools they employ, and to analyse the strategies undergraduate students employ in the thesis writing process. The study employed a qualitative approach where the primary method was a semi-structured interview. The interview included a Retrospective Think-Aloud (RTA) protocol where the participants shared their thoughts and strategies in completing the thesis. The participants were eight final-year undergraduate students majoring in Arabic and English language studies from a language faculty in a selected higher learning institution in the central region of Malaysia. Findings revealed four interconnected strategies – AI as Assistive Tool, AI as Intelligent System, AI as Cognitive and Affective Partner, and Delegated Autonomy. Mapped across the domains of AI literacy – Engaging, Creating, Managing, and Designing – these strategies represent a developmental continuum from mechanical aid to critically regulated collaboration. Among them, delegated autonomy emerged as a key theoretical contribution, describing students’ deliberate management of human–AI interdependence through ethical judgement and metacognitive control. These identified strategies could act as a framework to effectively use AI in thesis writing, contributing to the enhancement in the writing process and the development of students’ writing skills.