Briefing
Reference 1 · Primer · About 5 minutes
Download the MP3 · Linda, an AI voice, reads a script written from the sources below.
Written from public sources and checked on 10 October 2026. The guidance is paraphrased, not reproduced. Check the current text before relying on it.
The rules that apply to AI in Australia and for Australian organisations with European exposure, and the practical work each one calls for.
These are Agile Insights AI Strategy & Governance offerings. Which of them you need depends on where your organisation starts.
Sign up for our podcast at agile-insights.com.au.
Linda Australia has no AI Act, but it does have a national yardstick for responsible AI, and regulators will hold you to it. By the end of this episode, you will know the six practices, the law that already applies, and what to do first. This is the position as at October 2026.
Linda Here is what it is. The Guidance for AI Adoption was published on 21 October 2025 by the National AI Centre, which sits in the Department of Industry, Science and Resources. It sets out six essential practices for safe and responsible AI. It is voluntary and nobody certifies it. It replaced the ten guardrails of the Voluntary AI Safety Standard from 2024. It also comes in two versions. Foundations is for organisations starting out or using AI in low risk ways. Implementation practices is for complex, higher risk use.
Linda The problem it solves is a common one. Your people are already using AI, and the board asks, are we doing this properly? Australia has no single AI statute to point to. A bank might use a model to triage loan applications, and a supplier might embed AI in a product you buy. When something goes wrong, nobody can say who owned the decision. The guidance gives you a shared, plain description of what good looks like.
Linda Who uses it? Directors and chief executives ask for it in board papers. Risk and privacy leads use it to build policy. Procurement teams use it to question suppliers. Government agencies lean on related policy. And when a regulator asks what reasonable practice looked like, this is the document most likely to be put on the table. It is voluntary, but in practice it is the standard of care people will expect.
Linda Now the core. The six practices, in the published order. One, decide who is accountable. Name an owner for AI overall and for each system, including where a supplier is involved. Two, understand impacts and plan accordingly. Before you deploy, work out who is affected and how. Three, measure and manage risks, using risk management that is specific to AI. Four, share essential information. Keep a register of your AI systems, and tell people when they are dealing with AI. Five, test and monitor. Test before release, then keep watching, because behaviour can drift. Six, maintain human control. A trained person must be able to oversee the system, step in, and switch it off.
Linda How does this connect to the rest of the stack? Think of three layers. Data foundations sit underneath. In the middle sits AI governance, and the six practices belong there. They line up with management system standards such as I S O, I E C, 42001, so one programme can serve both. On top sits the law. The National AI Plan, released on 2 December 2025, chose to rely on existing technology neutral laws, with regulators responsible in their own areas. It also set up an AI Safety Institute to test and advise. It did not introduce a stand alone AI law, and commentators noted that the earlier mandatory guardrails idea was dropped.
Linda Those existing laws matter. The National AI Centre points to the Privacy Act, the Australian Consumer Law on misleading claims and outputs, discrimination law, work health and safety duties for physical and psychosocial risks, and directors' duties. From 10 December 2026, organisations covered by the Privacy Act must say in their privacy policies when computer programs make, or substantially help make, decisions that significantly affect people. The privacy regulator, the OAIC, issued its final guidance on that on 30 September 2026. Another useful group is SafeAI-Aus, an independent community resource with free templates. It is not government.
Linda Three first steps. First, build a register of every AI system and tool in use, including the ones staff adopted themselves, each with a named owner. Second, run a gap analysis against the six practices, using the Foundations version if you are early. Third, check your privacy policy and decision processes against the December obligation. Then read the primary sources: the National AI Centre guidance, the Department's pages, and the OAIC guidelines. The show notes link to them.
Linda Currency. As at October 2026, the guidance is current, and both versions were last updated on 5 May 2026. The Voluntary AI Safety Standard is retained for reference, but the department says the new guidance evolves it. On 15 July 2026, the Prime Minister announced an Office of AI and Australian Standards for AI, to be legislated early in 2027. The announced focus is large data centres and the training of AI on Australian creative works, not how businesses use AI. A joint parliamentary committee on AI is reportedly due to report on 30 November. These are proposals, so do not build against them yet. Always check the current text before you rely on any of this.
Linda To recap, six practices, existing law, and a record that proves you are doing what you say. If you'd like to know more, sign up for our podcast at agile-insights.com.au.
All of the reference series (16) · Follow AI Governance Reference in Apple Podcasts or any podcast app
Everything from Agile Insights: briefings, Making You Think and the reference series