PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
January 26, 2026British Journal of Educational Technology0 citationsOpen Access

Can Generative AI support the learning agency of students with disability? A case study of an Australian secondary school

View Full Paper
NRNatasha Anne RappaKNKaren P. NonisKTKok‐Sing Tang

Key Points

  • The study aims to explore how generative AI can enhance the learning agency of students with disabilities in a secondary educational setting.
  • Investigated the use of ChatGPT by students with disabilities in a science classroom.
  • Gathered data through student interviews and chat conversations, alongside lesson materials.
  • Conducted thematic analyses to identify learning choices and challenges faced by students.
  • Students expressed meaningful choices in using ChatGPT for their learning support.
  • Metacognitive challenges were identified, leading to struggles in aligning choices with learning capabilities.
  • Findings highlight the need to customize generative AI tools to better support the learning needs of students with disabilities.

Abstract

Abstract Despite increasing interest in using Generative Artificial Intelligence (GenAI) in education, little is known about how students with disability engage with GenAI to support their own learning. This study investigates the potential of ChatGPT to support the learning agency of adolescents with disability in a secondary science classroom in Australia. Guided by sociocultural and socio‐material conceptualisations of agency, the study explored the mediated choices and capabilities of three students with disability to use ChatGPT to facilitate their learning. The study was conducted in a class comprising students of varying ages clinically diagnosed with diverse learning needs. Data sources included student interviews, the students' conversations with ChatGPT, teachers' lesson worksheets and video recordings of the lesson. Thematic analyses reveal that while students expressed clear and meaningful choices to use ChatGPT to support their learning, they faced metacognitive challenges and cognitive constraints, resulting in a misalignment between their choices and actual capability. The findings identify key theoretical perspectives and practical considerations for supporting students with disability in using GenAI to develop their learning agency. The study recommends customising GenAI for specific learning needs in line with its function as a cognitive prosthesis for students with disability and for better alignment with Universal Design for Learning, thereby supporting students' learning agency. Practitioner notes What is already known about this topic The development of students' learning agency has been widely explored in the secondary context, but not for students with disability. There is increasing interest in using Generative AI (GenAI) to support inclusive education. Current theoretical frameworks can inform the exploration of the learning agency of students with disability using GenAI. What this paper adds Foregrounds the learning choices of students with disability as indicators of their emergent learning agency. Identifies learning challenges facing students with disability that misalign their learning choices and capabilities. Examines GenAI's potential as a socio‐material mediator facilitating the material and relational agency of students with disability. Implications for practice and/or policy Consulting students with disability on how GenAI can support their learning, providing opportunities for them to express their learning preferences. Customising GenAI tools in line with the Universal Design for Learning guidelines to address the specific learning challenges of students with disability. Clarifying the roles of teachers and education assistants in facilitating the learning agency of students with disability in light of customising GenAI to mediate their learning agency.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Rappa et al. (2026) studied this question.

synapsesocial.com/papers/697703d3722626c4468e8dd2https://doi.org/10.1111/bjet.70048
Ask AI
Helpful
Bookmark
Share
View Full Paper