Summary
The article critiques the Albanese government's pivot toward active AI regulation, shifting from a focus on productivity to a framework addressing social, moral, and sovereignty concerns. It highlights the establishment of a new Office for AI and legislated standards for data centers as evidence of a more cautious national stance. While acknowledging these governance steps, the author notes the government remains overly optimistic about AI being a controllable 'tool' similar to past technologies, potentially ignoring warnings from AI creators about the unpredictable nature of frontier systems. This development marks a significant shift in Australian AI policy, moving toward a more proactive, risk-aware regulatory posture.
Body
AdvertisementJuly 24, 2026 — 5:00amSaveYou have reached your maximum number of saved items.Remove items from your saved list to add more.ShareAAAThe Albanese government is finding its voice on technology. It has taken a curious path to get to this point, and on AI it has been led to it. But having been so led, it now celebrates leading the world. “Australia will be the first country in the world to bring these issues into a single, national framework,” the prime minister said last week, announcing new national standards for anyone wanting to build AI data centres here. These standards began life in March as environmental “expectations”, the government curiously refusing to give them the force of law. Now they will be legislated.This legislation is best seen as one of a pair, its sibling being the government’s teen social media ban. Albanese proudly boasts that “more than 20 nations have implemented or are implementing social media age restrictions of their own”, following his government’s lead. The success of the policy is, at this point, far from assured, but the government is clearly convinced its message resonates. It symbolises the idea that we have the power to leash the rapacious effects of the technology that has so quickly taken over our lives.Illustration: Simon Letch That itself is a watershed notion, a new public mood that is less inclined than previously to assume technology always makes life better. Social media benefited from that assumption, doing all kinds of damage along the way. The truth is, it’s probably too far gone. But “imagine if the world had acted a decade ago”, mused Albanese. “That is the opportunity – and the choice – we now have with artificial intelligence”. If so, it’s an opportunity the government had been choosing to ignore. Until now, it had sung in unison with the Productivity Commission, which hymned about AI’s productivity promises and recommended against any AI-specific regulation. Now the government wants to establish an Office for AI within the prime minister’s own department.Perhaps the most radical shift, though, is philosophical. Albanese’s speech made a point of trying to see AI not just in economic terms. He began with a rejection of the economics of Thatcher and Reagan, and especially Thatcher’s dictum that there is no such thing as society. He explicitly framed AI as a legal, social, and even moral and spiritual issue, citing the Pope’s AI encyclical along the way. The about-face is dramatic.AdvertisementThe change, I suspect, has three main drivers. In no particular order: the agitation of the Australian Council of Trade Unions ahead of the Labor Party Conference, which has been sounding the alarm about AI-caused job losses; the first signs of community discontent over planned data centres in places like the Blue Mountains; and One Nation’s attack on the government’s original AI policy, which Pauline Hanson described as deciding “largely to leave it in the hands of corporations with little regulation to ensure that the industry operates in the public interest”. All of these criticisms assess AI in terms that are not purely economic. They place questions of quality of life, and not merely wealth creation, at the centre of things. As such, they each capture perfectly a widespread anxiety the government had been avoiding.That’s what the Albanese government’s new approach is designed to address. The promised data centre standards are aimed mostly at energy and water use. They will require investors to contribute their own renewable power, and that the centres’ demands for resources will not cause energy prices to spike. The government wants Australia to develop its own AI as a way of retaining sovereignty – of shaping AI rather than being shaped by it. And the overarching promise is that by taking control now, we can ensure AI will deliver jobs, not take them. To that end, Albanese cites labour market research – “the first research of its kind by any government anywhere in the world” – showing graduate employment remains high and jobs in software and tech are growing.That’s encouraging news for the moment. But analysis from the US, which is further down the AI path than we are, is less encouraging. In April, Goldman Sachs estimated that AI was causing a net elimination of 16,000 jobs a month. Two months later, it reduced that figure to 11,000, but the difference was mostly down to the jobs being created in building data centres. These jobs are overwhelmingly short term, and once data centres are built, they require few staff to operate. As for the job losses, they’re hitting young people hardest. Goldman Sachs’ data is now starting to show that industries with higher AI adoption rates have higher unemployment among workers younger than 30. It’s an early trend, so it will need to be monitored further. And there will be plenty of opportunity to do so because the rate of AI adoption is only increasing.And this distils the main tension in the government’s new AI discourse. It talks a more social language, which is welcome. It claims to defend our non-economic goods, like sovereignty and the environment. But it also has to continue to court AI investment, including in the form of data centres, because it sees economic dividends too large to ignore. To resolve this tension, it simply must adopt a sunny view of what government policy can achieve. “We should not treat AI as a threat to good jobs,” declares Albanese. “We must use it as an instrument to help create them”, just as we have with technologies past. How, exactly, is not yet known.AdvertisementThe conceit here is that AI is a comparable technology to what has come before, controllable in a comparable way. The trouble is that so many of its creators say the opposite: that even they don’t fully understand how it works or what it will do. We’re yet to see if the government has any answer to that concern. For now, it seems not to share it. Perhaps it sees no alternative, knowing that for all the reassurances of sovereignty and control, AI’s future is mostly being written elsewhere.Whatever the case and for all the change in the government’s perspective, this optimism that AI remains a mere tool, an extension of the past rather than a departure from it, remains the government’s central, unchanging article of faith.Waleed Aly is a broadcaster, author, academic and regular columnist.The Opinion newsletter is a weekly wrap of views that will challenge, champion and inform your own. Sign up here.SaveYou have reached your maximum number of saved items.Remove items from your saved list to add more.ShareLicense this articleMore:AIOpinionFor subscribersAnthony AlbaneseALPWaleed Aly is a broadcaster, author, academic and regular columnist for The Age and The Sydney Morning Herald.AdvertisementAdvertisement