How to grade student projects fairly
Grading projects is hard because a polished submission does not prove the student understands it, and large classes make careful marking slow. A rubric, a short viva and a requirement checklist make grading both fairer and faster.
Write the rubric first
List four or five criteria, such as understanding, design, quality, problem solving and communication, and weight them. Share it with students so they know what counts.
Check the submission against the assignment
Turn your instructions into a checklist of requirements and mark each as met, partly met or missing, with a piece of evidence. This avoids grading on overall impression.
Add a short viva
Ten minutes of questions about the student's own work reveals whether they built and understand it. Ask questions from their actual code, data or report, and score each criterion from the answers.
Use AI as an assistant, not a judge
An AI assistant can ask consistent, project-specific questions, record transcripts and draft rubric scores for every student. You should review the scores, adjust where needed, and keep the final decision.
Be transparent
Tell students how they will be assessed, keep the transcripts so any score can be explained, and allow students to ask questions about their result.
Run vivas for your whole class
Set an assignment, share one link and let every student take an adaptive AI viva on their own project.
See how it worksFrequently asked questions
- How do I grade a project when I suspect it was not the student's own work?
- A short viva about the project is the most direct check: a student who built it can explain and justify it.
- Can AI grade student projects?
- It can draft scores against a rubric and save time, but teachers should review them, since no automated scoring is free of bias.