2026 ASEE Annual Conference & Exposition

Artificial Intelligence (AI) is rapidly transforming education, reshaping how students learn, study, and engage with complex material. While AI tools can enhance efficiency and provide quick access to knowledge, they often encourage surface-level learning when students use them merely to generate answers. This approach bypasses reflection and conceptual understanding, the very processes that enable critical thinking and problem-solving. As AI becomes a standard presence in higher education, there is an urgent need to rethink its role not as a substitute for learning, but as a catalyst for deeper engagement and cognitive growth.
This project investigates how AI can serve as a guided learning partner that helps students develop reasoning skills through structured, step-by-step problem-solving. Instead of providing direct solutions, the proposed system encourages students to think through each stage of a problem, promoting persistence, self-reflection, and meaningful learning. Drawing on theories of self-regulated learning and cognitive scaffolding, the system is designed to prompt students with targeted questions, explanations, and contextual hints aligned with course objectives.
The prototype integrates retrieval-augmented generation (RAG), generative AI, and classification models to deliver personalized, context-aware guidance. Using RAG, the system retrieves relevant course content and examples to ground the AI’s responses in the material taught by instructors. This ensures that feedback and hints remain pedagogically aligned, reducing hallucinations and preserving academic integrity. The architecture, developed with React and Next.js for the frontend and Python for the backend, supports interchangeable large language models (LLMs) and adaptive retrievers to maintain flexibility across disciplines.
The system is currently being implemented in a Discrete Mathematics course with approximately 300 to 350 students at a large public R1 university. A mixed-methods research design is being used to evaluate its impact on student engagement, confidence, and learning experiences. Quantitative data include system interaction logs and usage analytics, while qualitative data are gathered through surveys and reflective feedback. The analysis focuses on how guided AI interactions influence students’ motivation, problem-solving strategies, and self-regulated learning behaviors.
This work contributes to ongoing national conversations about the responsible integration of AI in education. It emphasizes that when AI is intentionally designed, it can complement instruction by reinforcing reasoning and metacognition rather than replacing human judgment. The long-term vision is to establish design principles for AI-powered educational systems that promote critical thinking, creativity, and lifelong learning, helping students become not just consumers of AI outputs but active participants in the learning process.