Date
Publisher
IEEE
Contributions: In this article, a generative artificial intelligence (AI)-based Q&A system has been developed by integrating information retrieval and natural language processing techniques, using course materials as a knowledge base and facilitating real-time student interaction through a chat interface. Background: The rise of advanced AI exemplified by ChatGPT developed by OpenAI, has sparked interest in its application within higher education. AI has the potential to reshape education delivery through chatbots and related tools, improving remote learning and mitigating challenges, such as student isolation and educator administrative burdens. Yet, ChatGPT practical applications in education remain uncertain, potentially due to its novel and enigmatic nature. Additionally, current e-learning chatbot systems often suffer from development complexity and a lack of input from key stakeholders, leading to developer-focused solutions rather than user-centered ones. Intended Outcomes: In this manuscript, we introduce a practical implementation of AI in education by creating a system called Q-Module-Bot that is accessible for both technical and nontechnical educators to harness e-learning benefits and demystify generative pretraining transformer (GPT). Application Design: The proposed Q-Module-Bot system has utilized pretrained large language models (LLMs) to build a Q&A system that helps students with their queries and supports education delivery using content extracted from a virtual learning environment (VLE). Findings: The prototype and system evaluation confirm the effectiveness of a scalable cross-departmental tool featuring source attribution and real-time responses. While successful in encouraging wider acceptance of GPT use cases in higher education, refinements are needed for full integration into the VLE and expansion to other modules/courses.
Study at a glance
- Question: How can a generative AI chatbot be designed to help students get answers from their course materials?
- Setting and data: Higher education, using course materials from a virtual learning environment.
- Method: Prototype development and system evaluation.
Why it matters
Practical applications of ChatGPT in education remain uncertain. Additionally, current e-learning chatbots often suffer from development complexity and a lack of stakeholder input, leading to developer-focused solutions rather than user-centered tools accessible to educators.
Key findings
- In higher education, an AI question-and-answer tool effectively provided real-time responses with source attribution across multiple departments.
- The tool successfully encouraged wider acceptance of generative AI use cases across higher education.
- The chatbot still requires refinements to achieve full integration into online learning platforms and expansion to other courses.
Summary drafted with AI from the paper’s abstract and may contain errors. See the full paper for details.
What is the application?
Who age?
Why use AI?
