The Prime Minister’s announcement of a new Office of AI inside the Department of the Prime Minister and Cabinet, together with a national framework for Australian Standards for AI, is more than another technology policy update.
It signals that AI is now core national infrastructure.
Not a side issue for innovation teams. Not a future-facing experiment. Not a narrow digital policy question.
AI now sits at the centre of national productivity, public trust, energy planning, data centre investment, creative rights, workforce capability and sovereign capability.
That matters because the AI debate has inverted.
A year ago, much of the public sector conversation was about whether government had enough internal capability to use AI safely and well.
Today, the Prime Minister’s announcement shows the question has moved.
It is no longer just: can government build enough capability to use AI?
It is now: can Australia set the rules, standards and trust settings that allow AI to operate in the national interest?
That is the right shift.
"This world leading framework is about Australia choosing to shape the future rather than letting the future of AI shape us." The Hon Anthony Albanese MP | Prime Minister of Australia
From capability gap to national standard-setter
At last year’s Government Innovation Week Federal in Canberra, Minister for Finance Senator Katy Gallagher placed AI inside the larger APS reform challenge.
Her message was not simply that government needed better tools. It was that government needed to rebuild the internal capability to understand, govern and deliver change.
She pointed to $20.8 billion in external labour, equivalent to a “shadow workforce” of more than 50,000 people, and warned that internal APS capability had eroded over the previous decade.
That was the capability question.
Could the APS rebuild the skills, confidence and operating discipline needed to manage change on its own terms?
Today’s announcement builds directly on that reform agenda.
The new Office of AI and proposed Australian Standards for AI are not just about regulating technology. They are about giving Australia a national operating model for AI: clearer accountability, stronger public capability, better infrastructure rules, safer adoption and more confidence for citizens, workers, creators, investors and agencies.
In simple terms: last year, the question was whether government could build the muscle to manage AI.
This year, Australia is starting to define the rules of the game.
AI is now an infrastructure issue
One of the most practical parts of the announcement is its focus on large data centres.
AI does not float in the cloud. It runs on real infrastructure: land, power, water, chips, networks, cooling systems and skilled people.
The Government’s proposed standards will require large data centres to underwrite their own new power supply, pay their full share of connection costs, reduce power when needed to support the grid, and be as water efficient as possible.
That is a serious and sensible move.
If AI is going to be treated as a national productivity opportunity, then the infrastructure behind AI must also be treated as a national planning issue.
Data centres should not push costs onto households. They should not weaken the grid. They should not ignore water pressure or local community concerns. They should not be approved through a patchwork of inconsistent rules.
AI infrastructure needs public trust as much as private capital.
That is why national consistency matters.
Investors need certainty. Communities need confidence. Governments need rules they can enforce. Agencies need assurance that the tools and systems they use are backed by responsible infrastructure.
This is where Australia can lead.
Not by being the loudest country on AI. Not by racing to the bottom. Not by pretending the risks do not exist.
Australia can lead by building one of the most trusted AI environments in the world.
Public trust will decide the speed of adoption
The Government’s language on social licence is important.
The Prime Minister said the framework is about Australia choosing to shape the future rather than letting the future of AI shape us.
That is the right frame.
AI adoption will not be decided by technology alone. It will be decided by trust.
If workers think AI is being used to weaken conditions, adoption will slow.
If citizens think AI is being used without transparency, trust will fall.
If artists, writers, journalists and creators see their work used without control or consent, the backlash will be real.
If communities believe data centres are draining local resources or increasing energy bills, social licence will disappear quickly.
That is why the Government’s commitment to protect Australian artists and media matters. The announcement makes clear that Australian writers, artists and journalists should retain ownership over their work, and that companies should not use Australian creative works to train AI without the artist’s control.
That principle is simple: innovation should not mean extraction.
AI systems depend on human knowledge, creativity and labour. If Australia wants a healthy AI economy, it must protect the people and institutions that create the material those systems rely on.
From AI hype to AI discipline
This is also what public sector leaders have been asking for.
At GIW Federal, the strongest feedback was not a call for more AI hype. Delegates valued the chance to discuss AI’s opportunities and challenges, but some also wanted more specific walkthroughs of innovation and clearer outcomes.
That is an important signal.
The market is maturing.
Public servants do not need another abstract AI conversation. They need practical guidance. They need standards. They need safe use cases. They need procurement models. They need assurance. They need workforce training. They need examples that show what works, what fails and what can be scaled.
The Prime Minister’s announcement responds to that shift.
It moves the national conversation from broad enthusiasm to practical discipline.
From pilots to standards.
From isolated agency activity to national coordination.
From AI as an innovation topic to AI as regulated economic infrastructure.
That is the inversion.
What this means for the public sector
For the public sector, this announcement should be treated as a starting gun.
Every agency will need to answer 6 practical questions:
- Where are we already using AI?
- Where could AI improve service delivery, productivity or decision support?
- What risks need stronger governance?
- What skills do our people need?
- What data can safely support AI use?
- How do we keep humans accountable for important decisions?
The Office of AI can help create national coordination.
But implementation will happen inside departments, agencies, councils and frontline teams.
That means the next phase must be practical.
Public servants will need clear playbooks, safe-use guidance, procurement rules, assurance models, workforce training and shared examples of what good looks like.
AI policy cannot stay at the level of principles. It has to become daily operating practice.
Australia cannot outsource its AI future
This is where the link to APS reform becomes critical.
Minister Gallagher’s earlier warning about eroded internal capability should stay front of mind.
Australia cannot outsource its AI future.
It can work with industry. It can attract investment. It can partner with universities. It can buy world-class technology. It can learn from global leaders.
But the public interest cannot be outsourced.
Government needs enough internal capability to ask the right questions, set the right standards, test vendor claims, protect citizens, govern data, manage risk and make decisions in Australia’s interests.
That is why the new Office of AI matters.
It creates a centre of gravity.
It gives AI a whole-of-government home.
It recognises that AI touches every minister, every department and every major part of the economy.[4]
And it sends a clear signal that AI is no longer just a technology conversation. It is an economic, social, infrastructure and governance conversation.
The leadership test
The real test now is execution...
Standards must be clear. Compliance must be practical. Infrastructure rules must be enforced. Public sector capability must be built. Creators must be protected. Communities must be heard. Workers must be trained. Agencies must be supported.
And the benefits must be visible to Australians.
If we get this right, AI can help lift productivity, improve services, strengthen sovereignty and support a more capable public sector.
If we get it wrong, AI will become another source of mistrust, fragmentation and uneven benefit.
The Prime Minister’s announcement gives Australia a stronger frame.
It says AI should work in Australia’s interests.
That is the right test.
Now the task is to make it real.
Want to learn more? Join us at Government Innovation Week in Canberra (21-22 Oct)
Help your peers
Share what you've learned with fellow public servants