What AI may do
- Extract text or structure from uploaded question papers, answer keys, and scripts.
- Suggest draft marks based on rubrics, sample answers, and marking instructions.
- Draft feedback comments for lecturer review.
- Flag uncertainty, missing answers, possible rubric conflicts, or submissions needing closer review.
What AI must not replace
- Final grading approval by a lecturer or authorized academic reviewer.
- Institution appeal, moderation, resit, disciplinary, or misconduct procedures.
- Professional judgment where a question requires interpretation, context, or academic discretion.
Accuracy and limitations
AI systems can misunderstand handwriting, notation, diagrams, ambiguous rubrics, partial-credit logic, or unusual student reasoning.
Lecturers should treat AI output as a draft and check it against the actual answer, rubric, marking guide, and course standard before publishing.
Student fairness
Institutions should provide a reasonable path for review, correction, or appeal where a student disputes a grade or feedback generated with AI assistance.
GradeOptimus should be used to support consistent grading, not to hide the basis of academic decisions.
Data use for AI
Student submissions and institution content should be used to provide the requested grading workflow and support the service.
GradeOptimus should not use identifiable student submissions to train public or unrelated AI models unless the institution has expressly agreed in writing.
Questions about this page?
These pages are practical product policies and trust summaries. Institution contracts can include more specific terms for deployment, data processing, support, and billing.