How AI Became a Study Partner in Statics

Faculty Spotlight | 2026 CTE AI Course Integration Grant

Most students use AI the obvious way: type the assignment into a prompt, copy the answer. Will Lindquist, teaching professor in Civil, Environmental, and Architectural Engineering, and doctoral student Marzieh Allahyari wanted to see how students could understand their own learning by introducing generative AI into practice problems.

Through the CTE AI Course Integration Grant, they designed a way to incorporate generative AI into their Statics course as something closer to a tutor that asks questions than a digital answer generator.

What they changed 

Lindquist and Allahyari didn't rebuild the course. Homework, lectures, topics, and exams all resembled their course design. The one new addition was a set of AI-supported practice assignments tied to specific exam problems, where students had to:

• Use AI to work through the problem setup and solution steps, but never to spit out the final number. 
• Keep reflection logs  where they questioned and checked what the AI told them. 
• Focus on catching their own mistakes before the exam, not after. 

This approach let them see the impact of generative AI by teaching students metacognition and error identification prior to taking exams.

How students responded

The move to introduce AI as a learning tool was a departure from how students typically use generative AI. “Most students responded positively to using AI as a structured learning tool,” Lindquist said. “We noticed right away, though, that many were not very skilled at using it effectively for learning.”

One unexpected outcome from incorporating AI was students learning how they learn. The AI couldn't help unless students showed their work clearly. The importance of writing more neatly and laying out their setups became clear to students. As one student explained, “The most helpful part of using AI was learning how to make my FBDs and problem setup clearer.… I can confidently say after doing these assignments, I know the importance of structuring my problems in a cohesive way.”

By the middle of the semester, students were using AI more carefully and trusting their own judgment about when it was wrong. As Lindquist explained, “With structured guidance early in the semester, students became more thoughtful in how they used AI and more confident in evaluating its responses.”

The results

• Scores on the AI-aligned exam questions rose by 1.51 points out of 10 (a difference-in-differences effect).
• Students reported real gains in confidence using AI and spotting bad answers (p<0.01).
• They had no AI on the exam itself, so those gains came from what they'd learned, not from the tool.

For Lindquist, the results emphasized the importance of designing AI activities around reasoning rather than answers. “The biggest lesson was that students need structure and coaching to use AI effectively for learning,” he said. “We designed activities that required them to analyze their setup, identify mistakes, question AI responses, and reflect on their reasoning rather than simply get an answer.”