Everyone
AI Use and Transparency
Where AI is used, when it is switched off, what is sent to the AI provider, and what it is never used for.
Effective · Last updated
1. Where AI can appear
| Feature | What the AI does | What it receives |
|---|---|---|
| Fictional colleagues | Writes an in-character reply to an email you send to a character in a mission | Your email, the character's role and the fictional mission facts the character may know |
| Coach | Offers a hint grounded in the mission's documents | Your question, the hint level, and the fictional mission material it may draw on |
| Mission authoring (instructors) | Drafts scenarios, documents and steps that the instructor then reviews and edits | The instructor's prompt and draft content — no learner data |
Coach answers produced by AI are marked “Generated guidance”. Characters are instructed never to grade you or reveal answers, and they cannot change the simulated world; only your own actions do that. We record whether each reply was generated or prepared in advance, so your Institution can check.
2. When AI is switched off
- By default, for every Institution, until it chooses to enable it.
- In assessment, independent-practice and demonstration modes, whatever the Institution's setting. AI can write only in the guided, assisted and scenario practice modes.
- When the Institution requires its data to stay in a particular region and the AI provider's declared region is not that region. The Service then uses prepared replies and makes no call to the provider.
- If the AI provider is unavailable: you get the prepared reply instead of an error.
3. Our AI provider
When AI is enabled, requests go to the provider named at Subprocessors. We send only what the feature needs (see the table above) — not your name or email address, unless you write them into your message. The provider processes it to produce the reply under its own terms for API customers; where those terms permit the provider to retain requests for abuse monitoring, that retention is the provider's and is described on the subprocessors page. Because the provider may process data outside the EU and Türkiye, Institutions should enable AI only when that is acceptable to them.
4. Limits of AI output
AI output can be wrong, inconsistent or incomplete. It is learning material inside a fictional world, not professional advice. It never decides a grade: evaluations come from the mission's defined criteria and from your instructors.
5. Academic integrity
Your Institution sets the rules for using outside AI tools on its courses. Follow them, and the Acceptable Use Policy. Do not try to make an AI character reveal assessment criteria or act outside its role.
6. Questions
Ask your Institution whether AI is enabled for your course, or write to us at privacy@itspracticing.com.