Skip to main content

How we use AI

Last updated: 10 July 2026.

The site has an optional AI study tutor ("Sentry") to help you understand the course material, look for the Ask Sentry button in the corner of the page. This page explains (in plain English) what it is, what it can and can't do, and how we keep it safe and accountable. It reflects the Australian Government's Voluntary AI Safety Standard, and we publish it so you can hold us to it.

1. What the AI tutor is​

The tutor is a chat assistant that explains cyber security concepts, re-explains them a different way when you're stuck, and quizzes you on the material. It runs on a large language model (Claude, made by Anthropic). It is an AI, not a person: it will tell you so if you ask.

Using it is completely optional. You never need the tutor to read the lessons, do the labs, or complete the course.

2. What it can help with​

  • Explaining a concept from the course in simpler terms or with an analogy.
  • Checking your understanding and quizzing you before an exam.
  • Pointing you to the right lesson or the official source for something.

It's a tutor, not an answer key. For graded quizzes and labs it will guide you toward the answer rather than just hand it over, because that's how you learn.

3. What it won't do​

  • It won't help with real-world attacks. It won't help you access, scan, exploit, phish, or surveil any system, account, or person you don't own and have written permission to test. Security concepts here are for defensive, authorised, educational use only, the same rule as our Terms. It can explain how an attack works and how to defend against it; it won't give you a usable procedure to harm a real target.
  • It won't give you advice to rely on. It doesn't provide legal, financial, migration, or careers advice you should act on without checking an official source.

4. It can be wrong: always verify important facts​

AI can sound confident and still be mistaken. Do not rely on the tutor for anything time-sensitive or consequential: course funding and eligibility, government programs, certification costs, exam details, salaries, or legal specifics. For those, check the official source we link or the course's own pages, which we keep current. If the tutor and a course page disagree, the course page wins, and please report it.

5. Your privacy when you use the tutor​

  • The messages you type into the tutor are sent to Anthropic to generate a reply. Don't type real passwords, credentials, personal details, or information about real systems. The tutor will remind you not to, and won't ask for them.
  • Your messages are used only, in the moment, to generate the reply. We don't keep your tutor conversations: there are no accounts, no chat history saved on our side, and nothing to log back into. We don't use your messages to train AI models, and we don't sell them.
  • Because nothing is stored, your conversation is gone once you close the tab. If you want to keep something the tutor told you, copy it yourself.
  • The rest of the site still collects nothing. This section applies only if and when you choose to use the tutor.

6. Human oversight and accountability​

  • The project maintainer is personally accountable for the tutor and will review how it's performing.
  • We test the tutor adversarially (a "red-team" suite). It passed before launch, and we re-run it before any change, checking that it refuses harmful requests and defers on facts it shouldn't guess.
  • We can switch the tutor off instantly if something goes wrong.
  • The tutor never makes a decision about you: it only explains. It can't grade you, gate your progress, or change your access.

7. Tell us if it gets something wrong​

If the tutor gives you a wrong, harmful, or inappropriate answer, please report it. Every report is read by the accountable owner and used to improve the tutor. Because the tutor is optional, you can also simply choose not to use it. The whole course works without it.

8. Questions​

Questions about how we use AI? See Report a concern.