Sri Lanka Institute of Information Technology (SLIIT) Malabe, Sri Lanka, to bridge the gap between evolving learning styles and the power of artificial intelligence, a collaborative research team from three Sri Lankan universities has developed “RAG SLIITbot,” a specialised AI chatbot designed to provide students with accurate and contextually relevant academic support. The innovative tool aims to address the challenges posed by general-purpose AI models like ChatGPT, which, despite their popularity, can sometimes deliver lengthy, inaccurate, or inconsistent responses when used for studying.
The project, a joint effort between Sri Lanka Institute of Information Technology (SLIIT) Malabe, SLIIT Northern Uni Jaffna, and the University of Peradeniya, focuses on creating an AI assistant that leverages document-based information retrieval and generation. RAG SLIITbot is built on a sophisticated Retrieve and Generate (RAG) architecture, allowing the Lecturer-in-Charge (LIC) to upload DOCX, PDF, and TXT files containing lecture notes, exam papers, and other essential learning materials. The chatbot then pulls information directly from these sources to answer student queries. Plans are in place to expand functionality to include future PowerPoint (PPT) files.
The Need for a Customised AI Solution
We recognise students’ increasing reliance on AI for quick and efficient learning. However, general AI models don’t always align with a particular module’s specific content and teaching style. RAG SLIITbot is designed to solve this problem by providing responses exclusively based on the academic documents offered by lecturers.
A key differentiator of RAG SLIITbot is its integration of Natural Language Processing (NLP) and vector databases like Pinecone. This allows the chatbot to deliver accurate and contextually relevant answers, avoiding the pitfalls of irrelevant or misleading information that can sometimes plague general AI models. Furthermore, the system efficiently analyses images and charts within uploaded documents using AI-based image transcription techniques. This enables visual data extraction, including diagrams, figures, and graphs, to enhance response generation. While some advanced LLMs like Gemini and ChatGPT also offer image analysis capabilities, this feature enhances the system’s ability to process and integrate visual content effectively. This feature is particularly significant, offering students a richer and more comprehensive learning experience. The lecturer added that by extracting information from text and visuals, RAG SLIITbot ensures students receive a holistic understanding of their study materials by chatting with RAG SLIITbot.
The development of RAG SLIITbot highlights the increasing importance of customised learning systems in the age of AI. Unlike generic AI models, these tailored tools ensure that information is precisely aligned with the curriculum and educators’ unique teaching styles. This enhances student engagement and promotes academic integrity by encouraging reliance on verified sources.
Furthermore, the chatbot’s ability to adapt to individual learning styles is expected to be a significant advantage. RAG SLIITbot is designed to provide concise, subject-specific answers derived only from the prescribed academic materials, making it particularly useful for exam preparation. Beyond its practical applications, the project also plays a crucial role in documenting the methodologies and workflows in building document-based AI systems. This knowledge base is critical in shaping the future of AI-powered education and ensuring that institutions invest in tools that support meaningful, student-centred learning.
Security is also a paramount concern. To protect sensitive academic data, uploaded documents undergo anonymisation during processing, ensuring that private academic materials remain secure. With its emphasis on accuracy, relevance, and security, RAG SLIITbot represents a significant step forward in integrating AI into Sri Lanka’s higher education landscape. The tool promises to empower students with a powerful and efficient learning assistant tailored to their academic needs. Research and development into tailored academic AI tools like RAG SLIITbot will be essential in building a more innovative and future-ready education system as AI evolves.
The Future of AI-Powered Education
These customised AI solutions will encourage students with powerful, efficient learning assistants designed to meet their academic needs. Projects like RAG SLIITbot will contribute significantly to shaping the future of AI-powered education. RAG SLIITbot stands out from general AI chatbots due to its document-based information retrieval approach. It ensures responses are precise, short, and strictly based on Lecturer-in-Charge (LIC) -uploaded materials like lecture notes and exam papers. Unlike general chatbots, such as ChatGPT, which provide lengthy and sometimes inconsistent answers, RAG SLIITbot extracts data only from verified academic sources, ensuring accuracy and relevance. Its Natural Language Processing (NLP) and vector database integration allow it to deliver concise, subject-specific responses, making it an ideal academic assistant tailored for Sri Lankan higher education.
In conclusion, the advent of RAG SLIITbot marks an essential milestone in integrating AI into Sri Lanka’s higher education system. By offering a tailored and secure AI solution that adapts to individual learning styles and ensures academic integrity, RAG SLIITbot promises to improve the learning experience for students in the country and contribute to the growth of AI-powered education globally.
Moreover, future iterations of RAG SLIITbot could incorporate multilingual support, enabling students from diverse linguistic backgrounds to access academic content in their preferred language. Additionally, integrating speech-to-text capabilities could assist visually impaired students and those who prefer auditory learning. The chatbot’s potential expansion to include interactive learning features like quizzes, flashcards, and real-time collaboration tools would further enhance student engagement. By continuously evolving to meet the dynamic needs of learners, RAG SLIITbot exemplifies how AI can be harnessed to create a more inclusive, adaptive, and student-centric educational ecosystem.
Interactive and informative chatbots like RAG SLIITbot enhance the learning experience by providing students with clear, well-structured, and relevant information. Such studies significantly improve students’ understanding and foster innovation in education. If Sri Lankan educational institutions and individuals across the country actively work towards developing their own AI-powered academic tools, it would create a more personalised and effective learning ecosystem.
AI-driven chatbots can offer students better insights, assist in research, and support innovative learning strategies.
Furthermore, universities and research institutions should focus on expanding research and development in AI-driven educational tools, ensuring they align with the curriculum and evolving learning needs. Investing in these technologies will enhance digital literacy in Sri Lanka and position the nation as a leader in AI-integrated education. By embracing Generative AI (GenAI) in education, Sri Lanka can bridge the digital divide, improve accessibility, and create a future-ready academic environment. These advancements will strengthen the nation’s education system, preparing students to thrive in a technology-driven world and contributing to national progress.
“Technology will never replace great teachers, but in the hands of great teachers, it can be transformative.” – George Couros
Prof. Roshan G. Ragel.
Department of Computer Engineering,
University of Peradeniya, Sri Lanka.
Email: [email protected]
Prof. Nuwan Kodagoda.
Pro Vice-Chancellor, Faculty of Computing,
Sri Lanka Institute of Information Technology (SLIIT), Sri Lanka.
Email: [email protected]
Mr. Kanagasabai Thiruthanigesan, Lecturer,
Faculty of Computing,
Sri Lanka Institute of Information Technology (SLIIT) Northern Uni, Jaffna, Sri Lanka.
Email: [email protected]
Mr. Kithusshand Raveendran,
Academic Instructor,
Faculty of Computing,
Sri Lanka Institute of Information Technology (SLIIT) Northern Uni, Jaffna, Sri Lanka.
Email: [email protected]
Ms.Gowrisha Karuneswaran,
Temporary Assistant Lecturer,
Faculty of Computing,
Sri Lanka Institute of Information Technology (SLIIT) Northern Uni, Jaffna, Sri Lanka.
Email: [email protected]
Mr. Vasavan Parthipan, Temporary Assistant Lecturer,
Faculty of Computing,
Sri Lanka Institute of Information Technology (SLIIT) Northern Uni, Jaffna, Sri Lanka.
Email: [email protected]
