BoilerTAI: A Platform for Enhancing Instruction Using Generative AI in Educational Forums
Project Overview
The document explores the use of BoilerTAI, a platform that employs Generative AI (GenAI) to enhance online educational forums by assisting instructional staff in addressing student inquiries more effectively. By pre-answering questions, GenAI significantly reduces the workload of Teaching Assistants (TAs), enabling them to manage larger classes more efficiently. Grounded in Vygotsky’s socio-cultural theory, the study positions AI as a 'More Knowledgeable Other' (MKO) that supports learning. Findings indicate that students respond positively to AI-generated answers, which are of high quality; however, the importance of human oversight is underscored to ensure the accuracy and relevance of the information provided. Overall, the implementation of GenAI in education demonstrates promising potential in optimizing instructional support and enhancing the learning experience for students.
Key Applications
BoilerTAI - a platform integrating Generative AI with online educational forums.
Context: First and second-year computer programming courses at Purdue University.
Implementation: The platform was introduced in a controlled experiment, using AI-TAs to respond to student queries on discussion boards.
Outcomes: Improved efficiency and quality of responses, reduced cognitive load for TAs, and positive student reception.
Challenges: Need for human oversight to ensure accuracy and context; potential for incorrect or misleading AI responses.
Implementation Barriers
Perception Barrier
Instructors display caution regarding AI's accuracy, privacy, and ethical implications, while students generally have a positive view of AI technologies.
Proposed Solutions: Training for instructors on AI capabilities, emphasizing the human-in-the-loop model to mitigate concerns.
Technical Barrier
AI-generated responses may lack context and evidence, leading to potential inaccuracies. Human oversight is integrated into the response approval process to ensure accuracy and relevance.
Proposed Solutions: Ensure that human oversight is part of the response approval process to validate the accuracy and context of AI-generated content.
Project Team
Anvit Sinha
Researcher
Shruti Goyal
Researcher
Zachary Sy
Researcher
Rhianna Kuperus
Researcher
Ethan Dickey
Researcher
Andres Bejarano
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Anvit Sinha, Shruti Goyal, Zachary Sy, Rhianna Kuperus, Ethan Dickey, Andres Bejarano
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai