Abstract:

Disabled students in Higher Education (HE) persistently face wide-ranging and complex physical and psychological barriers to accessing and learning from information presented via a pervasive teaching practice of verbal delivery accompanied by textual and visual content. Disabled students are more likely to drop out than their non-disabled peers and barriers, such information processing and concentration, contribute to systemically lower levels of attainment. Developments in AI and GenAI can be leveraged to address this inequity: accurate Automatic Speech Recognition (ASR) and ethical applications of GenAI to create learning materials offer powerful ways to improve the learner experience. Truly personalised learning materials, including study skills support personalised to the learner, are now possible and could transform the learning journey of disabled students.

Messenger Pigeon (MP) is an app that creates transcripts and learning materials from HE content such as lectures and slides for use by disabled learners. MP utilises ASR, text-to-speech and GenAI in its Study Assistant and can be provided to disabled students as an accommodation to support access to HE lectures. Personalised learning materials will be more effective if they are tuned to the psychological profile of the learner. To achieve this goal the MP development team are working with psychologists to apply psychological theory, research and methodologies to create a framework of learning behaviours that could be supported in MP to personalise the learning experience of disabled learners.

The research involved 3 stages:
(1) Literature review which identified effective learner behaviours, such as self-regulation of learning, motivation, and self-testing while highlighting that technological (AI) acceptance and psychological readiness were also key
(2) User experience research involving audio-diaries and qualitative interviews (n=5). Participants reported a cautious attitude towards the use of AI in education and expressed preference for a global approach integrating MP use with other study skills including note-taking and self-testing
(3) Quantitative survey collection to establish learning profiles of existing MP users.

The Learning and Study Skills Inventory (LASSI) was circulated to active MP users (n=1000). LASSI measures 10 aspects for successful learning which are compared against a large database of answers to produce percentile scores for respondents. 48 MP users completed LASSI. Learner profiles were highly informative in identifying areas of low learner skills: with 48% of students scoring in the 1 percentile range for one or more behaviour and 23% scoring in the 1 percentile for 3 or more behaviours. Results indicate that MP users need extra support to be able to engage with their learning through the development of learning behaviours, skills and strategies. By developing a nuanced understanding of an individual learner’s profile these can be targeted through AI Study Assistant prompt development to personalise the learner journey.

Unlike traditional learning analytics which track behaviour, these findings offer insight to potential causes of the behaviour such as high levels of anxiety and low levels of motivation.This collaborative research highlights the importance of psychologically grounding technological developments to personalised learning strategies and materials and identifies key areas for future research and development of AI to support inclusion.

Keywords:

AI, disabled students, neurodivergent students, inclusion, higher education.