AI Usage Policy

Last updated: 15 October 2026

Current status

Ocevalon has one optional AI feature, for examiners only: a suggested mark for a student script, produced by Google Gemini. It is off unless the site administrator turns it on. Ocevalon does not send your posts, messages or profile to any AI service.

How the examiner suggestion works

  • An examiner must press a button for each paper. Nothing is sent automatically.
  • The question text, the student’s typed answers and the uploaded answer-page images for that one paper are sent to Google Gemini.
  • The AI returns suggested marks and short reasons. These are not saved. The examiner reads the pages and decides every mark.
  • Handwriting and equations can be misread. The AI is told to say when something is hard to read, but it can still be wrong. Marks are never finalised or published by the AI.
  • We keep a record that a suggestion was requested (who and when), not the AI text.

Reading a scanned question paper

An examiner who creates an exam can upload a scanned or photographed question paper (PDF or image) and press a button to have Google Gemini read the questions into the exam form. The file is sent to Google only when the button is pressed. The examiner must check every question before saving. Do not upload papers that contain personal information about students.

If more AI features are added

Before any AI feature launches, we will update this policy and the Privacy Policy to explain:

  • which features use AI and which provider processes the data;
  • what you submit, and what you must not submit (for example private, confidential or exam material);
  • how long data is kept, based on the provider's actual terms;
  • any usage limits.

Using AI responsibly

  • AI output can be wrong or incomplete. Check important facts and references.
  • Do not use AI to cheat, impersonate, harass or create illegal content.