Australia Levels Up: A New Era for AI Governance 🇦🇺

The National AI Centre’s Guidance for AI Adoption streamlines last year’s Voluntary AI Safety Standard’s (VAISS) 10 guardrails (published in September 2024) into 6 essential practices and, crucially, turns them into actionable steps for teams at different maturity levels. This is more than just a rebrand; it's due to active listening. It reflects extensive 2024–25 consultation and evidence that industry needed clearer, more actionable guidance they could tailor to context.

What changed — and why it matters?

• From 10 guardrails to 6 practices:

• Cleaner structure, less duplication, and practical “how-to” actions for developers and deployers.

• Two guided entry points. Start where you are:

• Foundations: for teams new to AI or using lower-risk cases.

• Implementation Practices: for organisations scaling AI, handling higher-risk use, or needing deeper governance detail.

• Built-in tools:

• AI screening tool, AI policy template, and AI register template help you stand up the basics quickly.

• Aligned, not overwhelming:

• Stays consistent with Australia’s AI Ethics Principles and international approaches—while making compliance achievable.

Side-by-side: Which version should you use?

The 6 essential practices — what to do at each level?

Both guides use the same six practices. Foundations shows the baseline actions, and Implementation Practices adds the advanced version when risk or scale increases.

Where we can help

We transform the Guidance into operations, establishing policies and registers, running the screening tool, right-sizing risk and testing, embedding human-in-the-loop controls, crafting transparency statements, and preparing audit-ready evidence. Whether you’re piloting your first assistant or scaling across the enterprise, we tailor Foundations vs Implementation to your context.