HomeAbstract of 11th-ntteconfDesigning AI for Learning, Not Just Answers: A Dialogic Support System for Positive Change in Learning Behaviours

Designing AI for Learning, Not Just Answers: A Dialogic Support System for Positive Change in Learning Behaviours

  • Aug 2, 2026
  • Posted by:
  • Category: Abstract of 11th-ntteconf
No Comments


Abstract Book of the 11th International Conference on New Trends in Teaching and Education

Year: 2026

[PDF]

Designing AI for Learning, Not Just Answers: A Dialogic Support System for Positive Change in Learning Behaviours

Dr. Nazim Ali, Rabia Areej, Sarah Aynsley, Keira Kirkland Dowler-Morley, Fabio Fedi, Raina Lele

ABSTRACT:

Assuming all students start university with equal learning skills can lead to academic expectations that disadvantage certain student groups and undermine their confidence. Sector reports have highlighted that gaps in learning skills are key drivers of attainment disparities and that, alongside curriculum reform, structured interventions are needed to help students strengthen learning behaviours. In response, this study examined whether a custom made and curriculum aligned GenAI dialogic learning system could support positive change in students’ learning behaviours. Rather than functioning as a generic chatbot, the system was designed to act as a subject specialist tutor to guide students to think and respond through scaffolded, structured dialogue. Working with students from Global Ethnic Majority (GEM) backgrounds who experience attainment disparities, the system was co-designed through a needs analysis and evaluated through pre- and post-intervention surveys, supported by student feedback. Developed using an AI agent-based process, the system provided students with step-by-step questioning, helped them identify gaps in understanding, and offered scaffolded support for tackling complex concepts. Students reported positive changes in learning behaviours, including a shift from rote learning towards more active learning, self-monitoring, and application of knowledge. Rather than revising through passive review alone, students reported using the system to diagnose weak areas and structure their learning more effectively. We therefore argue that, to move beyond AI’s use as an answer generator and reduce the bypassing of critical thinking, AI should be integrated into learning environments in ways that support scaffolded learning and foster positive change in students’ learning behaviours.

Keywords: Attainment; Equity; Learning; Skills; Scaffolding