The Academic Integrity module adds originality checking directly into your assessment workflow. Instead of a standalone plagiarism percentage, it compares submissions against your own institutional repository and flags similarity for staff to review. This guide gets you from an empty repository to your first originality report.
Important: A similarity score is an indicator, not a verdict. Final academic-misconduct decisions always remain with authorised staff under your institution’s policy. The module never applies a sanction on its own.
What it checks (Phase 1)
The native engine compares each submission against material you already hold:
- Other students in the same cohort — the strongest signal for collusion.
- Prior students’ submissions across earlier cohorts.
- Your institutional repository of accepted work and exemplars.
- Exact duplicates — identical files or text, detected instantly.
Phase 1 does not search the public internet and does not use AI-writing detection — those are later phases by design, so results stay explainable and defensible.
Where to find it
Open the sidebar under Academic Integrity:
- Integrity Dashboard — live counts: submissions checked, average similarity, high-risk, exact duplicates, repository size, open cases.
- Originality Checks — every check, filterable by type, risk band, and assessment.
- Similarity Heatmap — student-to-student matches for one assessment.
- Repository — the corpus used for comparison, with exclude/delete controls.
- Integrity Reports — PDF/Excel/CSV exports.
How a submission gets checked
Whenever a learner submits an exam free-text answer, uploads evidence, or submits an assignment, an originality check runs in the background and its result appears under Originality Checks. Each checked submission is also added to the repository, so detection improves the more your institution uses it.
You can also trigger a check directly via the API (POST /api/integrity/checks) with the submission text.
Reading a report
Open any check to see its originality report:
- Overall similarity and a risk band — Low, Review, or High.
- A breakdown: same-cohort student, any prior student, and institutional repository.
- Matched sources with the reference and match percentage.
- A count of highlighted overlapping passages.
Risk bands come from sensible defaults per assessment type (an essay flags earlier than a technical assignment, which naturally reuses terminology). These thresholds become configurable per programme in a later phase.
Next steps
- Learn the review-and-escalation flow in Handling a plagiarism case.
- Encourage learners to use the Originality Self-Check in their portal before submitting.