TBR AI EXCHANGE
Submission Date
Submitter’s Name/Email
Institution/School
Department/Discipline
Activity Purpose (assessment, data collection, classroom management, etc.)
AI Tool(s)
Activity Details
As Co-Principal Investigator on the National Science Foundation (NSF) project Excellence in Research: Artificial Intelligence-Driven Learning Analytics to Predict Student Performance in STEM Education at Historically Black Colleges and Universities (Award #2503017), I sought to improve student engagement and provide more personalized learning support in Architectural Design I and II, and Architectural Graphics. Traditional classroom instruction often makes it difficult to identify struggling students early enough for timely intervention. The project aimed to leverage AI-driven learning analytics to provide instructors with actionable insights while offering students adaptive learning opportunities outside the classroom. During the first year of implementation, the project served 104 undergraduate STEM students across four courses.
I integrated the Knowt AI learning platform into my Architectural Design and Graphics courses as a supplemental instructional tool. Students completed AI-generated practice sets that reinforced course concepts and design principles outside of class. The platform collected learning analytics including practice frequency, time-on-task, completion rates, and response accuracy, allowing me to monitor student engagement and identify students who would benefit from additional instructional support. These analytics informed classroom discussions, targeted feedback, and individualized interventions. I also used ChatGPT and Copilot to assist with lesson planning, instructional materials, and developing reflective learning activities that encouraged students to critically evaluate AI-generated information rather than simply accepting AI responses.
The implementation demonstrated strong early success. According to the independent external evaluation, students responded positively to AI-supported learning tools and reported that AI-generated practice activities helped reinforce course content. Learning analytics showed sustained student engagement beyond initial access, and instructors reported increased confidence in using AI-generated analytics to guide instructional decisions. The project also established a robust data collection infrastructure that supports future development of AI-driven prediction models for student success. One important lesson learned was that AI literacy must accompany AI adoption. Students require explicit instruction on validating AI-generated information, understanding AI limitations, and maintaining academic integrity. These insights have led me to redesign assignments to include reflection, verification, and ethical AI use as integral components of instruction.