- Posted on 28 Jul 2026
- 4 min read
The project investigated the experiences of students with learning disabilities and used those insights to develop GuideLight Chat, a prototype generative AI chatbot.
Supporting students with learning disabilities to navigate university using AI
For many university students, planning their degree can be confusing. There are subjects to choose from, prerequisites to understand and complex rules that determine how a course is completed.
For students with learning disabilities, these challenges can be even greater.
Patrick Djizmedjian, a student researcher in the UTS Faculty of Engineering and Information Technology, set out to better understand these barriers and explore whether generative AI could help make university information easier to access and understand.
Supported by a UTS Social Impact Grant, Patrick worked with supervisors Dr Ali Braytee and Dr Lucy Bryant to investigate how students with learning disabilities experience the UTS Handbook and to develop a practical solution that could improve access to critical academic information.
The result was GuideLight Chat, a prototype AI-powered chatbot designed to help students navigate handbook information through plain-language explanations while maintaining links to official university sources.
Turning a complex problem into an accessible solution
The project focused on students with learning disabilities, including ADHD, dyslexia and autism spectrum conditions.
These students often encounter barriers when navigating large volumes of information presented in dense academic language. While the UTS Handbook is the authoritative source for course and subject requirements, understanding and interpreting its content can be challenging.
Patrick's research found that many students experience cognitive overload when trying to locate and interpret essential information.
Some described the process as overwhelming.
One student reported there was ‘too much to click through, too many redirects,’ while another reflected that it had ‘taken me a pretty long time to figure this stuff out myself.’
For Patrick, these responses reinforced the importance of finding new ways to improve accessibility.
"The project enabled us to move beyond identifying a problem and begin developing a practical solution," he says.
"We wanted to understand the specific challenges students face and build something that reflected those needs."
The team conducted an online needs analysis survey with students to better understand their experiences. Participants highlighted shared challenges such as interpreting prerequisite requirements, locating relevant information and managing large amounts of text.
These findings became the foundation for the development of GuideLight Chat.
Designing AI with accessibility in mind
Rather than simply providing AI-generated answers, the team took a careful approach to building trust and transparency into the system.
GuideLight Chat was designed to retrieve information directly from UTS Handbook content and present it in shorter, structured explanations that are easier to digest. Importantly, the tool also provides links back to official handbook sources so students can verify information themselves.
The approach addresses a key concern when using AI in educational settings.
"Trust calibration is essential," Patrick says.
"Students want clear and structured support, but they also need simple ways to check information against official sources."
The project brought together expertise from computer science and speech pathology, helping ensure that both technical and accessibility considerations informed the design process.
This interdisciplinary collaboration was critical to creating a tool that responds to the needs identified by students themselves.
As Patrick explains, "The strongest message from participants was that navigation friction and understanding progression requirements were major barriers. Those insights directly shaped the design priorities for the prototype."
Creating a more inclusive university experience
While the project is still progressing, it has already delivered important outcomes.
The research produced a detailed understanding of how students with learning disabilities experience institutional information and identified the support features they value most. It also resulted in a functioning prototype ready for further testing and refinement.
The next stage of the project will involve evaluating the chatbot with students and gathering feedback on its usability, accessibility and effectiveness.
The long-term goal is to explore how technologies like GuideLight Chat might contribute to a more equitable learning environment.
"This work provides a foundation for ongoing research into accessibility-oriented generative AI in higher education," Patrick says.
"There is real potential to reduce barriers, improve access to information and help students navigate university with greater confidence and independence."
The project also points to broader opportunities beyond the UTS Handbook. Feedback from participants suggested that navigating administrative processes, policies and forms can present similar challenges, highlighting future possibilities for expanding accessibility-focused support tools across the university.
At its heart, the project is about ensuring that all students can access information they need to succeed.
By combining student insight, interdisciplinary expertise and emerging technology, the research demonstrates how small innovations can contribute to a more inclusive higher education system.
With thanks to:
UTS partners
Dr Ali Braytee and Dr Lucy Bryant
The grant application
This project was supported by a 2025 UTS Social Impact Grant.
Webpage
Hidden Disabilities Sunflower Initiative at UTS
At UTS, we are proud to join the Hidden Disabilities Sunflower initiative, a global movement that supports people living with non-visible disabilities.
Webpage
Hidden Disabilities Sunflower Initiative at UTS
At UTS, we are proud to join the Hidden Disabilities Sunflower initiative, a global movement that supports people living with non-visible disabilities.