On 15 July, in a keynote titled AI in Australia's Interests at the University of Sydney, Prime Minister Anthony Albanese did something more consequential than announcing a new coordinating office. He ended five years of voluntary AI governance in Australia outright, describing AI as a bigger challenge and a bigger opportunity than social media, and committed the government to a package of Australian Standards for AI that will, for the first time, carry enforceable obligations rather than opt in expectations.
That is a meaningfully different statement to the one most organisations have been preparing for. Voluntary guidance can be adopted at a pace an organisation chooses. Enforceable standards cannot.
What "Mandatory" Actually Changes
Australia has had AI governance guidance for years. What it has not had, until this announcement, is a requirement that organisations comply with it rather than simply consider it. The distinction matters because most Australian AI governance activity to date has been shaped by voluntary frameworks that organisations could adapt selectively, adopting the parts that suited their existing processes and quietly deprioritising the parts that did not.
Enforceable Australian Standards for AI remove that flexibility. Once a standard is mandatory, partial adoption is not really compliance, it is a documented gap. Organisations that have been treating AI governance as a maturity journey they control the pace of are now looking at a deadline they do not.
The Two Standout Elements Everyone Is Missing
Two specific elements of the framework deserve more attention than the coordinating office announcement they arrived alongside. The first is a mandatory rulebook covering large data centres and AI training itself, extending governance obligations upstream to the infrastructure AI is built on, not just the applications built with it. The second is a requirement that Australian creative works cannot be used to train AI systems without the artist's control, including control over the price and value of that use.
That second element in particular is a genuinely new position. Most AI training data debates globally have focused on whether use is permitted at all. Australia's framework goes further, giving creators commercial control over the terms of that use, which is a materially different and more assertive stance than most jurisdictions have taken so far.
Why the Readiness Gap Makes This More Urgent Than It Looks
The timing is not especially convenient for the market it applies to. Only 22 percent of Australian organisations say they are mostly or fully ready in AI governance, compared with 33 percent globally, even as AI now supports 29 percent of tasks in the average Australian business, up from 25 percent a year earlier. Adoption has been accelerating faster than readiness, and mandatory standards are arriving into exactly that gap.
This is where the announcement's practical weight becomes clearer than the coordinating office alone suggested. A voluntary framework arriving into a readiness gap is a recommendation to catch up. A mandatory one arriving into the same gap is a compliance deadline for roughly three quarters of the market.
How This Connects to the Office of AI Already Coordinating the Detail
The Office of AI, established the same day inside the Department of the Prime Minister and Cabinet, is the body actually designing these standards in detail, working across ministries to bring the data centre, training and creator consent elements into a coherent whole of government framework. The practical elements organisations should already be building around, a use case register, named accountable ownership, pre deployment assessment and incident reporting, remain the same regardless of exactly how the Office's standards are eventually worded. What changes with this announcement is the certainty that those standards will be mandatory rather than advisory.
What Organisations Need to Do Before the Standards Are Finalised
The organisations best positioned when the Australian Standards for AI are finalised will be the ones that used the gap between announcement and enforcement to close their own readiness deficit, rather than waiting for the exact text. That means an honest assessment of where the organisation sits against the 22 percent already governance ready, a clear view of where AI training data and large scale infrastructure obligations might apply, and a use case register built now rather than after a compliance deadline makes it urgent.
What This Means for Your Organisation
What we see across the organisations we work with is that the gap between voluntary and mandatory governance is where the real work usually gets done, or does not. Organisations that used similar transition periods overseas to build real infrastructure, rather than waiting for a final rulebook, consistently found compliance considerably easier once the mandatory deadline actually arrived.
Key Takeaways
- On 15 July, Australia ended five years of voluntary AI governance, committing to enforceable Australian Standards for AI rather than opt in guidance.
- Two standout elements of the new framework are a mandatory rulebook for large data centres and AI training, and a requirement that Australian creative works cannot be used for AI training without the artist's control over price and value.
- Only 22 percent of Australian organisations say they are mostly or fully ready in AI governance, compared with 33 percent globally, even as AI already supports 29 percent of tasks in the average Australian business.
- The Office of AI, established the same day, is coordinating the detailed design of these standards, but the core governance infrastructure organisations need does not depend on the exact final wording.
- Organisations that build a use case register, named accountable ownership and an assessment process now will find mandatory compliance considerably easier than those waiting for the final standard.
How Trusenta Can Help
AI Governance builds the use case register and accountable ownership structure that any version of the Australian Standards for AI is likely to require.
Compliance Management tracks obligations across the new mandatory standards alongside existing frameworks in a single view as the detail is finalised.
AI Governance Enterprise closes the readiness gap this post describes for organisations that need governance infrastructure built quickly ahead of enforcement.
Conclusion
Australia has spent years treating AI governance as something organisations could adopt at their own pace. That changed on 15 July. The specific wording of the Australian Standards for AI is still being designed, but the direction is no longer in question, and the organisations that spend the gap before finalisation building real governance infrastructure will not be the ones scrambling when enforcement actually begins.
