In July 2026 ASQA released its Principles for the Responsible Use of Artificial Intelligence in VET.
They are there to help providers use AI safely and ethically in training and assessment, and to keep quality outcomes while they do it. ASQA has been clear: the principles do not introduce new regulatory requirements. They are a structured way to apply the obligations you already have, including the 2025 Standards for RTOs, in an environment where AI is in the mix.
Read the live page on asqa.gov.au (search for Responsible use of Artificial Intelligence (AI) in VET). This article is a working translation, not a substitute for that guidance, and it is not legal advice.
The five statements ASQA asks providers to ensure are:
- AI use is supported by strong governance that ensures it does not undermine the quality or integrity of VET.
- Human oversight and accountability are maintained in all AI-supported activities, ensuring that decisions affecting students remain the responsibility of qualified trainers, assessors and staff.
- AI systems and tools manage information securely and in accordance with existing privacy, data protection and record keeping obligations.
- AI use supports and enhances student equity, inclusivity, accessibility and wellbeing.
- AI use aligns with training product requirements, industry expectations and the needs of the relevant student cohort.
1. Governance That Protects Quality and Integrity
Someone owns AI. It is not a collection of private ChatGPT logins on trainer laptops.
Know which tools are in use, who approved them, and what job they do. If a tool would weaken assessment integrity, student records, or the credibility of a qualification, it does not belong in the college, no matter how fast it is.
A simple register is enough to start: tool, purpose, owner, data it can see, review date.
2. Human Oversight and Accountability
“The AI decided” is not an answer you can take to a student, an employer, or an auditor.
AI can draft, suggest, flag, and give students practice. A qualified person still makes the call that affects the student: competence, support, enrolment facts, anything that commits the RTO.
Name the human for each AI-supported step. Write it into the procedure. Keep a trail of what was drafted and what was approved.
3. Secure Information, Privacy, and Records
This is the principle many providers find hardest, and it is the one that fails quietly.
Do not paste student files, identity documents, or funding details into a consumer tool to “see what it says.” Use systems you have chosen, with access limited to people who already handle that work. Keep the official record in your student management system, not in a chat history.
If you would not put it in a shared inbox email, do not put it in a prompt.
4. Equity, Inclusion, Accessibility, and Wellbeing
AI should help more students, not only the digitally confident ones.
A tool that mishears accents, ignores disability access, or has no alternative pathway can lock people out of demonstrating competence. Test with the cohort who will struggle first. Keep a human option. Watch whether outcomes split unfairly after you introduce a tool.
Principle 4 is not a slogan. It is whether every student can still show what they can do.
5. Alignment with the Training Product, Industry, and the Cohort
The unit of competency still wins. Industry still has to recognise the skill. The student still has to be able to do the job in a workplace, not only inside a simulation.
If AI helps someone produce a polished answer that does not match how the work is done on site, you have a validity problem. Use AI to support practice and admin. Keep assessment evidence authentic, sufficient, and mapped to the product you are delivering.
ASQA’s own case material makes the same point: an AI avatar can be useful for rehearsal and still be the wrong sole source of assessment evidence.
How to Use the Five Without Building a Second Rulebook
Do not write an “AI policy” that sits in a drawer next to the Standards. Fold the questions into the governance, assessment, privacy, and student-support processes you already run.
A short self-check before you switch something on:
- Who approved this tool, and what job does it do?
- Who is accountable if the output is wrong?
- What student or staff information does it see?
- Can every relevant student use it fairly, or is there another way?
- Does this still assess the training product, not the student’s ability to game the tool?
If those answers are clear, you are using the principles as ASQA intended: existing obligations, applied where AI actually shows up.
For the official wording, case studies, and self-assurance questions, use ASQA’s guidance on asqa.gov.au.
