After years of consultation papers, expert panels and voluntary standards, the Australian government has now put its intentions beyond doubt. Prime Minister Anthony Albanese has confirmed that Australia will enact laws to govern artificial intelligence, according to reporting carried by The Manila Times. The statement marks a clear shift in tone from a government that had, until recently, leaned heavily on encouraging industry to regulate itself.
From voluntary guardrails to actual law
Australia’s approach to artificial intelligence has to this point rested on a patchwork of existing rules, sector-specific obligations and a set of voluntary standards released to nudge companies towards safer deployment. The government floated a range of options over the past two years, including mandatory guardrails for so-called high-risk uses of the technology, but stopped short of committing to a dedicated statute. That hesitation drew criticism from unions, academics and civil society groups who argued that voluntary codes carry no consequences when they are ignored.
By confirming that legislation is now on the table, the Prime Minister has effectively conceded that the light-touch settings were never going to be enough on their own. The detail of what any law will actually cover remains the open question, and it is the detail that will decide whether the move satisfies the technology industry, worried workers, or neither. What is clear is that the government intends to give itself statutory tools rather than relying solely on persuasion and existing consumer, privacy and anti-discrimination law.
Two views on whether laws help or hinder
Supporters of a legislated regime argue that clear rules are good for everybody, including the companies building and selling the technology. Certainty about what is permitted, they say, lets businesses invest with confidence instead of second-guessing where the government might eventually land. Consumer advocates add that Australians deserve enforceable protections when automated systems make decisions about their jobs, their loans, their healthcare and their interactions with government. Without a law that bites, they argue, there is little to stop a company deploying a flawed model and simply apologising afterwards.
The counter-view comes largely from parts of the technology sector and from economists who worry about Australia falling behind. Their concern is that a heavy compliance burden could deter investment, slow the adoption that drives productivity, and push talent and capital offshore to jurisdictions with friendlier settings. Some point to the European Union’s sweeping AI Act as a cautionary tale, warning that Australia risks importing complexity without importing the scale that makes such rules workable. The challenge for Canberra is to write something that protects people without smothering a young industry, and reasonable people disagree sharply about where that line sits.
What it means for Australia
For Australian businesses, the practical upshot is that the era of treating AI governance as optional is drawing to a close. Banks, insurers, retailers and government agencies that have been racing to embed generative tools into their operations will need to watch closely, because a legislated regime is likely to impose transparency, accountability and testing obligations on the highest-stakes uses. The major banks in particular have been vocal about putting AI to work across fraud detection, customer service and lending, and they will want to know early whether their existing controls will pass muster.
There is also a sovereignty dimension. Australia buys most of its foundational AI capability from a handful of overseas providers, which means any domestic law has to grapple with how to hold offshore developers to account for products used locally. That is a hard problem, and it is one the European Union and others are still working through. For local startups, the stakes cut both ways. Clear rules could level the playing field against larger overseas rivals, but poorly calibrated compliance costs could fall hardest on the smaller players least able to absorb them.
Workers and their representatives will be watching the employment angle most closely. The debate over AI in Australia has increasingly centred on jobs, from job cuts attributed partly to automation through to concerns about surveillance and automated management. Whether the coming legislation addresses workplace uses of AI directly, or leaves that to the industrial relations system, will shape how the union movement responds. Expect the Australian Council of Trade Unions and individual unions to press for explicit protections rather than principles.
The politics of getting it right
The announcement lands in a charged political environment. State governments have already begun staking out their own positions, with debate in New South Wales and Victoria over protecting jobs and public trust as AI spreads through the economy. A federal law would sit above that activity and, ideally, harmonise it, though the history of Australian federalism suggests coordination will not be simple. The government has also been at pains to counter what it has framed as overblown claims from parts of the technology industry, insisting that some of the sector’s promises do not stack up.
Getting the design right matters because AI regulation is genuinely difficult. Define the scope too narrowly and the law is obsolete before the ink dries. Define it too broadly and ordinary software risks being swept up in obligations meant for frontier systems. Most experts favour a risk-based model that reserves the strictest requirements for uses with the greatest potential to cause harm, such as those affecting safety, legal rights or access to essential services. Whether Australia adopts that framing, or borrows more heavily from another model, will become clearer as draft provisions emerge.
What happens next
The immediate task for the government is to translate a stated intention into an actual bill, and that process will involve further consultation with industry, civil society and the states. Expect exposure drafts, submissions and no shortage of lobbying before anything reaches Parliament. Businesses would be wise to use the interval to map where they are relying on automated decision-making, document their testing and oversight, and prepare for a world in which those practices may need to withstand external scrutiny. The direction of travel is now set. The detail, and the fight over it, is only beginning.
Sources: The Manila Times.
