In the Australian Vocational Education and Training (VET) sector, AI use by students and assessors must align with the principles of assessment(validity, reliability, flexibility, fairness) and the rules of evidence (validity, sufficiency, authenticity, currency). These are outlined in the Outcome Standards (2025) and enforced by bodies like the Australian Skills Quality Authority (ASQA). Breaches can undermine qualification integrity, lead to non-compliance, and affect learner outcomes.
AI behaviours can introduce risks such as inauthentic evidence, biased outcomes, outdated content, or unfair processes. To address this, I’ve developed a risk matrix as a practical tool for RTOs, trainers, and compliance teams. It categorises common AI behaviours by students and assessors, evaluates them against the principles/rules, and assigns risk levels based on:
- Likelihood: How probable the behaviour is to occur or cause issues (Low: Rare/unlikely; Medium: Possible; High: Likely/common).
- Impact: Severity of consequences on assessment integrity, compliance, or learner equity (Low: Minimal disruption; Medium: Moderate issues, fixable with oversight; High: Significant breaches, potential regulatory action or qualification invalidation).
- Overall Risk Level: Calculated as a product of likelihood and impact (Low: Green; Medium: Yellow; High: Red; Extreme: Black). This follows standard risk management frameworks (e.g., AS/NZS ISO 31000).
The matrix we’ve created is presented in tables for clarity. One for students and one for assessors. Risks are informed by VET guidelines, including ASQA’s focus on academic integrity, requirements for human oversight in AI use, and ethical considerations like bias and transparency. RTOs should adapt this matrix with context-specific data (e.g., via audits) and implement controls like AI detection tools, clear policies, and training.
Student AI Behaviours Risk Matrix
Assessor AI Behaviours Risk Matrix
This matrix can be used by entering specific scenarios and scoring them dynamically (e.g., in a spreadsheet). For example, if a behaviour’s likelihood increases due to new AI tools, recalculate the risk. RTOs should integrate this into their compliance frameworks, drawing from resources like ASQA’s transparency statements and VET-specific AI guidelines to ensure responsible adoption. If AI use is disclosed and overseen, many risks can be reduced to low levels, promoting innovation while safeguarding standards.
Check out our standard AI and Plagiarism policy notes.
