It is one of the more seductive lines in the artificial intelligence debate: if machines can read, write, summarise and decide faster than people, surely they can do the same for the paperwork that clogs government. That is roughly the argument put by Sky News host Chris Kenny, who used a recent segment to suggest the AI revolution should help to “minimise” bureaucracy across the public sector.
It is a tidy proposition, and a politically potent one. Australians of most persuasions have a complaint about red tape, whether it is the wait for a development approval, the forms attached to an aged care package, or the compliance load carried by small business. Framing AI as the tool that finally clears the backlog taps a deep well of frustration. Whether the technology can actually deliver on that promise, and at what cost, is a far more contested question.
The pitch
The case Kenny gestures at is not fanciful on its face. Large language models are genuinely good at the sort of work that fills government in-trays: drafting correspondence, triaging requests, checking documents against rules, and pulling answers out of dense policy manuals. A public servant who spends half a day summarising submissions or reconciling a spreadsheet is doing exactly the kind of task these tools were built to accelerate.
The productivity numbers being spruiked around the sector are not trivial either. Governments here and overseas have run pilots where officials save meaningful chunks of their week on routine drafting and research. If you believe the bureaucracy is bloated, the logic runs, then a technology that lets fewer people do more of the processing work should let you shrink it. On Kenny’s telling, as reported by Sky News Australia, AI is less a threat to government than an opportunity to strip out its excess.
It is worth being clear that this is a commentary position, not a policy announcement. It sits alongside a broader push on the political right, in Australia and elsewhere, to treat AI as a lever for a smaller, cheaper state.
The complication
Here is where the neat theory meets the messy practice. “Minimising bureaucracy” and “automating bureaucracy” are not the same thing, and conflating them is where a lot of these arguments come unstuck. Much of what looks like red tape from the outside is, on closer inspection, an accountability mechanism: a record that a decision was made, by whom, on what basis, and with what right of review. Speeding those steps up is one thing. Removing the human who can be questioned about them is another.
Australia already has a cautionary tale on exactly this point. The Robodebt scheme, which used automated income averaging to raise unlawful debts against welfare recipients, was the subject of a royal commission that found the automation of a government function without adequate human judgement caused real and lasting harm. Any pitch to let algorithms take over decision-making in the public service now runs straight into that memory. The lesson was not “automation is evil”, but rather that automating a process you do not fully understand, at scale, against vulnerable people, is a recipe for disaster.
There is also a workforce dimension. Public sector unions have been vocal that AI in government should augment staff rather than serve as a quiet justification for cuts. Their concern is that “minimise bureaucracy” becomes a euphemism for headcount reduction, with the savings booked upfront and the risks discovered later. The counter from efficiency advocates is that a public service freed from rote processing could redirect people toward the frontline work that citizens actually notice. Both things can be true, and which one wins depends almost entirely on how the rollout is governed.
What it means for Australia
This is not an abstract debate here. The federal government has stood up an Office of AI and is working through a national framework for how the technology is used, including inside the Australian Public Service. Agencies have run trials of AI assistants for staff, and the Digital Transformation Agency has issued guidance on safe adoption. The direction of travel is toward AI in government, not away from it. The open question is the shape.
For Australia specifically, three constraints will decide whether Kenny’s vision translates into anything real. The first is trust: survey after survey shows Australians are more sceptical of AI than people in comparable economies, and they are especially wary of it making decisions about their money, health or entitlements. A government that automates too aggressively risks a backlash that sets the whole agenda back. The second is capability: much of the public sector still runs on ageing systems and fragmented data, and you cannot bolt a modern AI layer onto records that do not talk to each other. The third is oversight: without clear rules on when a human must be in the loop, and a genuine right to review an automated decision, efficiency gains can curdle into the next Robodebt very quickly.
There is a federalism wrinkle too. Bureaucracy in Australia is not one thing. It spans Commonwealth departments, state agencies, and local councils, each with its own systems, workforce agreements and appetite for risk. A commentator can talk about “the bureaucracy” as a single blob. Anyone trying to reform it has to deal with dozens of them, each moving at its own pace.
What is next
Expect the rhetoric to keep running ahead of the reality for a while yet. The political appeal of AI as a red-tape cutter is obvious, and it will feature in plenty of speeches before it features in many delivered savings. The more useful signals will be quieter: which agencies publish evaluations of their AI pilots, whether those evaluations measure quality and error rates rather than just time saved, and whether the government’s emerging framework puts hard limits around automated decision-making that affects people’s rights.
Kenny’s underlying instinct, that a lot of government process is heavier than it needs to be, is one plenty of Australians share. The hard part was never identifying the problem. It is doing the surgery without nicking an artery. AI makes that surgery faster, which cuts both ways: faster to fix, and faster to break. The bureaucracy that emerges on the other side will be defined less by the technology than by the guardrails put around it.
Sources: Sky News Australia.
