Course Review Assistance: A Human-Guided Approach
You built the course, so you know what every assignment is supposed to do. That familiarity also makes drift hard to spot: think the outdated announcement, a module copied forward, or an activity that no longer lines up with its objective. Instructure worked with Idaho State University and the University of Iowa to test whether AI-assisted course review could catch those gaps. This whitepaper shares what they found, including where the AI helped and where faculty still made the call.
What the course review exploration found
The team built a course design rubric from established frameworks like Quality Matters (QM) and OSCQR. Then they ran a prototype on courses in Canvas by Instructure test environments, with instructional designers and faculty reviewing the reports.
| What the team tested | What they found |
|---|---|
| Can AI apply an established rubric? | Yes. Reports lined up with criteria for course structure, organization, and alignment. |
| Can it spot course drift? | It flagged outdated announcements, inconsistencies across modules, and changes that built up over semesters. |
| Does faculty want automated verdicts? | No. They wanted to see the reasoning and evidence behind each item. |
| Did faculty act on every item? | No. They weighed each one against their own instructional intent first. |
| Where does it fit best? | Instructor self-review, faculty and instructional designer conversations, and revisiting courses that have changed over time. |
Download the whitepaper to see why collaborators valued the prototype as a second set of eyes on a course, with the educator deciding what changes.
Ebooks & Buyer's Guides
