Scaffold or shortcut?

Postgraduate IT students’ use of generative AI and self-regulated learning

Abstract

Generative artificial intelligence (GenAI) tools such as ChatGPT and Copilot are increasingly integrated into higher education, where students use them to summarise texts, solve problems, and generate code. While these tools are can reduce cognitive load and improve learning efficiency, they may also challenge students’ ability to regulate their learning (i.e., self-regulated learning; SRL) by encouraging surface-level engagement and overdependence. This study investigates how GenAI shapes SRL behaviours within a postgraduate information technology (IT) subject/unit/course. A mixed-methods design was employed with 267 students, combining pre- and post-semester surveys with semi-structured interviews. The study examined how students engaged with GenAI and how this affected SRL components of goal setting, monitoring, and self-evaluation. Findings show varied patterns: some students used GenAI to clarify goals, check understanding, and reflect on progress, while others relied on it as a shortcut, outsourcing monitoring and evaluation. The study highlights GenAI’s dual role as a scaffold and shortcut, offering insights for designing learning environments that foster productive use and sustain student agency and autonomy.