Every government eventually meets a technology it cannot ignore and cannot easily control. For the Albanese government, artificial intelligence has become exactly that problem, and the political challenge is no longer whether to act but how far to lean in each direction. Push too hard on regulation and Australia risks scaring off investment and falling behind trading partners already racing to build the technology. Push too softly and the government wears the blame the first time an AI system denies someone a loan, misreads a medical scan or floods the internet with convincing fakes.
That tension sits at the heart of an ABC News analysis of the government’s balancing act on AI, which frames Canberra as caught between two competing instincts. One says slow down and build trust before the technology outruns the public. The other says speed up before the economic window closes.
How Canberra got here
The debate has been building for a couple of years. After the launch of generative tools such as ChatGPT put AI in front of ordinary Australians, the then industry minister Ed Husic ran a consultation on safe and responsible AI, and the government signalled it favoured mandatory guardrails for AI used in high-risk settings, things such as hiring, credit, health and law enforcement. The logic was straightforward enough. Most everyday uses of AI, from spam filters to spellcheckers, need no special oversight. The harm sits at the pointy end, where an automated decision can reshape a person’s life without them ever knowing an algorithm was involved.
Since then the political weather has shifted. Following the 2025 election, responsibility for the industry and science portfolio moved to Tim Ayres, and the government’s language has tilted noticeably towards productivity and growth. Treasurer Jim Chalmers has repeatedly framed AI as a central lever for lifting Australia’s stubbornly weak productivity numbers, a message that landed hard at the government’s economic reform roundtable. The subtext is not subtle: a new law that slows AI adoption is a harder sell in a room full of business leaders than it was during a wave of public anxiety about deepfakes and job losses.
Two views pulling in opposite directions
On one side sits industry, and much of it wants Canberra to resist the urge to write AI-specific legislation. The Productivity Commission has been influential here. In its work on harnessing data and digital technology, the Commission cautioned against rushing to create a bespoke AI Act, arguing that a thicket of new, technology-specific rules could duplicate existing privacy, consumer and anti-discrimination protections while adding cost and uncertainty for the firms Australia most wants to attract. Tech companies and their peak bodies have echoed that line, pointing out that many AI harms are already illegal regardless of whether a machine or a human caused them.
On the other side sit unions, digital rights groups and a good chunk of the public, who argue that existing law was never designed for systems that make opaque decisions at scale. Their case is that voluntary principles and industry goodwill have a poor track record, and that without enforceable obligations, transparency about when AI is being used, avenues to contest an automated decision, and testing for bias, ordinary people carry all the risk while companies capture the upside. Robodebt hangs over this argument like a warning. That scheme was not artificial intelligence, but it showed how badly automated government decision-making can go wrong, and how long it can take for the damage to surface.
The government’s difficulty is that both sides are partly right. Overreach really could deter the data centre operators, model developers and enterprise adopters now weighing Australia against other markets. Underreach really could leave the country exposed when, not if, a high-profile AI failure hits the headlines.
Why this matters for Australia
For a mid-sized economy that mostly buys its frontier technology from the United States and increasingly from China, the stakes are less about building the next great model and more about setting the terms on which that technology is used here. Australia has form as a rule-maker that punches above its weight, from the news media bargaining code to online safety laws that other countries have studied closely. A clear, workable AI framework could be a genuine competitive advantage, giving businesses and citizens confidence that automated systems operating in Australia meet a baseline standard.
There is also a sovereignty dimension the government keeps returning to. Billions of dollars are flowing into Australian AI infrastructure, from Firmus and its backers building large-scale AI factories to the major banks standing up their own frontier capabilities. Much of the underlying technology, though, is controlled offshore. Getting the regulatory settings right is part of how Canberra tries to ensure Australians are not simply price-takers on tools that will increasingly mediate work, money and public services. Get it wrong in either direction and the country either forfeits investment or forfeits control.
The productivity argument cuts close to home too. Modelling floated during the reform debate has suggested AI could add meaningfully to national output over the coming decade, but only if adoption is broad rather than confined to a handful of big firms. That reframes regulation as an enabler rather than a handbrake. Clear rules can give a cautious small business the confidence to deploy AI, whereas a legal grey zone tends to freeze exactly the mainstream adoption the government says it wants.
What happens next
The likely path is incremental rather than a single sweeping statute. Expect the government to lean first on existing regulators, privacy, consumer protection, corporate and online safety watchdogs, to stretch current law over AI, while reserving mandatory guardrails for the highest-risk uses. A review of the Privacy Act, updates to online safety settings and any move on automated decision-making in government will all serve as proxies for how ambitious Canberra is really prepared to be.
What is clear is that the balancing act will not resolve itself. Every month the technology advances faster than the policy cycle can move, and the political cost of hesitation compounds alongside the cost of overreach. The government has bought itself time by talking about productivity and promising proportionate rules. Sooner or later it will have to show which way it leans when the two instincts finally collide.
Sources: ABC News.
