How do I check if a student used AI to write their work?
Viva does not scan text for AI. It asks the student to explain their own work, which is a much harder thing to fake.
Text detectors guess from style, get it wrong often, and fall apart the moment a student rewrites a paragraph. Viva takes the other route entirely: it asks the author questions only the author can answer well.
Alongside the interview, the Submission analysis reads what the FILE says about itself: Content Credentials (C2PA) manifests that AI tools embed, document properties naming the application that made it, and generator tags inside images. These render only when a mark is actually present. Absence proves nothing, because marks strip on copy-paste, so no mark is never treated as evidence of human writing.
What the interview surfaces:
- Whether the student can account for the specific choices in their own submission — why this method, why this source, why this number.
- Whether they can go one step past what is written, which is where borrowed work runs out.
- Whether their explanation contradicts what they handed in.
- Whether they are reading an answer aloud rather than speaking it.
Viva never tells you a student cheated. It shows you what they could and could not explain, in their own words, and you decide. That distinction is what makes the evidence usable in an academic integrity process.
Instructors get the rest — the walkthroughs, the screenshots, and Ask Viva, which finds the right page from a typed question.