Facilitating Personalized Learning with Real-time Support in STEM Using Large Language Models
Jose Torres, Andre Bouvier, Zhenyu Lin
School of Engineering
Faculty Supervisor: Zhuwei Qin
Improving personalized learning is key to boosting student engagement and performance in STEM education, but it remains challenging in large classrooms. Additionally, underrepresented minority (URM) students and women face barriers such as stereotype threat and lack of support, making traditional academic help less accessible. While AI tutoring platforms like ChatGPT and Claude offer scalable, adaptive learning solutions, concerns over misinformation, bias, lack of transparency, academic integrity, and affordability limit their effectiveness. To address these issues, the proposed project introduces AICO (AI Collaborative Educational Platform for Personalized Learning). AICO is a transparent, customizable, and open-source AI tutoring system designed to support critical thinking and independent learning. It uses course-specific materials to ensure accuracy and reduce bias, and faculty can refine AI outputs, improving model reliability. The platform supports multimodal learning with features like lecture summaries, interactive homework help, flashcards, and video and audio explanations. AICO integrates with Canvas and allows faculty to track progress and generate assessments. The project will be evaluated in three phases, involving student feedback, course deployment, and benchmarking against commercial AI models. Ultimately, AICO aims to create equitable, ethical, and student-centered AI-driven STEM education.