2026 ASEE Annual Conference & Exposition

The rapid evolution of generative artificial intelligence (AI) presents both opportunities and challenges for engineering education. While existing AI tutoring systems offer instant feedback and personalized resources, they raise concerns related to transparency, instructor oversight, student over-reliance, and equitable access. To address these challenges, the AI Collaborative (AICO) platform was developed at San Francisco State University, a primary undergraduate and hispanic serving Institution as an open-source, student-centered system that integrates course-specific instructional materials with large language models (LLMs). AICO prioritizes transparency, critical engagement, and instructor agency by grounding AI responses exclusively in instructor-curated content and emphasizing reasoning-focused support rather than direct answer provision.
AICO was piloted during Summer 2025 in a junior-level Mechanics of Solids course, a core requirement in civil and mechanical engineering curricula. Instructors uploaded textbooks, lecture slides, and problem sets into the platform, enabling AICO to generate interactive practice problems, step-by-step solution guidance, and personalized previews and reviews of course topics. Students engaged with AICO through conversational interactions designed to provide adaptive scaffolding, targeted hints, and course-aligned content navigation.
To examine student experiences, post-course reflection questions were administered, including Likert-scale items assessing perceived helpfulness, reliability, satisfaction, and likelihood of recommendation, along with open-ended prompts addressing benefits, limitations, accessibility, and suggestions for improvement. Platform usage logs provided additional contextual insight into how students interacted with the platform. Findings indicate that while many students perceived pedagogical value in AICO’s reasoning-focused support, technical performance, accessibility, and usability strongly shaped adoption and overall perceptions. This pilot highlights both the potential and the implementation challenges of instructor-guided, retrieval-augmented AI platforms in high-challenge engineering courses.