Researchers piloted a generative AI tool that gives electrical engineering students feedback on practice problems in real time, instead of making them wait on a human demonstrator.
The tool was tested in a large electrical engineering class, where the ratio of students to teaching staff is typically too high for anyone to get fast, personalized help. Students worked through practice problems on the platform and got scaffolded feedback immediately. Humans still mattered, but their role shifted from grading individual answers in real time to verifying the platform's model solutions ahead of each session. The researchers then studied how students actually used the tool and how much they trusted what it told them.
Feedback works best when it arrives fast, but fast feedback usually means nobody checked it first. This pilot tests a middle path: verify the answer key up front, then let the AI handle the live back-and-forth, which could make quick, trustworthy feedback workable in classes where staffing can't keep pace with demand.
It's a modest fix for a problem every large engineering program runs into: too few humans, too many stuck students, and feedback that shows up too late to change how anyone studies.