Australia’s simmering debate over how to govern artificial intelligence has boiled over again, with a Greens senator accusing the federal government of standing aside while the technology reshapes the economy, the workforce and the information Australians consume every day.
In comments reported by The Guardian, the senator argued Australia is “sleepwalking” into an AI crisis and warned the current settings amount to a “tech bro free-for-all” in which powerful overseas platforms write the rules and everyone else lives with the consequences. The phrasing is deliberately provocative, but it lands on a genuine policy gap: nearly three years after generative AI went mainstream, Australia still has no dedicated, enforceable AI law.
The context
Successive governments have leaned on a light-touch approach. The Albanese government released a voluntary AI Safety Standard in 2024 and ran a consultation on mandatory guardrails for AI in “high-risk” settings, covering areas such as testing, transparency and human oversight. But those guardrails remain proposals rather than legislation, and much of the day-to-day governance of AI in Australia still falls back on existing laws covering privacy, consumer protection, discrimination and copyright, none of which were written with large language models in mind.
That vacuum is what the Greens are targeting. The party’s broad position has been that voluntary codes and industry goodwill are no substitute for a binding legal framework with real penalties. Their concern spans several fronts at once: automated decision-making in areas like welfare, insurance and hiring; the scraping of Australians’ personal data and creative work to train foreign models; the spread of AI-generated misinformation; and the concentration of enormous power in a handful of US firms.
What the senator is arguing
The senator’s core criticism, as reported, is one of pace and priorities. On this view, Canberra is moving at the speed of consultation papers while the technology moves at the speed of product launches, and the mismatch leaves ordinary Australians exposed. The “tech bro” framing is pointed at the culture of a small number of dominant developers, the argument being that Australia has effectively outsourced its digital future to companies that answer to shareholders in California rather than to Australian law.
The Greens are also mindful of the labour dimension. As AI tools spread into white-collar work, from customer service to legal drafting to media, the party has pushed for guarantees around job security, retraining and the right of workers to know when a machine is making or shaping decisions about them. Where the Greens want the rules to land is unambiguous: a standalone AI Act with mandatory, enforceable obligations, an empowered regulator, transparency requirements for high-risk systems, and stronger privacy and copyright protections so that Australians’ data and creative output cannot be harvested without consent or compensation.
David Shoebridge, the Greens’ spokesperson on digital rights, has been among the party’s most vocal voices on these questions, consistently arguing that self-regulation has failed in other tech domains and will fail again here. The party frames the coming term as a narrow window: get the settings right now, or spend the next decade trying to claw back power already handed away.
The government and industry response
The government’s counter is that it is neither asleep nor idle, but deliberately careful. Its position, broadly, is that regulation should be risk-based and proportionate, tightening the rules on genuinely dangerous or high-stakes uses of AI while avoiding a blanket regime that could smother a technology the Treasury sees as a major productivity lever. Officials point to the existing consultations, the voluntary standard and reviews of privacy and copyright law as evidence that reform is under way, just methodically.
Industry groups have pushed in the same direction, with an added warning about competitiveness. Bodies such as the Tech Council of Australia have argued that heavy-handed, EU-style rules risk driving investment offshore and slowing the adoption of tools that could lift wages and output. Their preferred model is to build on existing laws, use sector regulators where possible, and reserve hard rules for clearly high-risk applications. Between these camps sits a large group of academics, civil society organisations and legal experts who agree that action is overdue but differ on how prescriptive it should be.
What it means for Australia
The stakes here are distinctly local. Australia is overwhelmingly a consumer of AI rather than a builder of frontier models, which means the systems shaping Australian hiring decisions, insurance quotes, news feeds and government services are largely designed and trained overseas. That dependence sharpens the regulatory question: without domestic rules, Australians are effectively governed by decisions made in boardrooms they cannot vote out.
There are concrete pressure points. Australian newsrooms, artists and musicians are watching the copyright fight closely, having seen their work vacuumed into training datasets with little recourse. Robodebt remains a live warning about what happens when automated decision-making goes wrong at scale, and it hangs over every conversation about AI in the public service. And with a federal budget under pressure, the productivity promise of AI is politically irresistible, which is precisely why critics fear safeguards will keep losing to speed.
What’s next
The immediate battleground is legislation. The government must decide whether to convert its proposed high-risk guardrails into a bill, fold AI obligations into a modernised Privacy Act, or continue with the patchwork of voluntary standards and existing laws. The Greens, along with a bloc of crossbenchers, will use their Senate leverage to press for something firmer, and are likely to keep the “sleepwalking” language front and centre to force the issue onto the agenda.
Expect the debate to intensify as more Australians encounter AI in their workplaces and public services, and as international benchmarks, particularly the EU’s AI Act and moves in the United Kingdom and United States, give local advocates a yardstick to measure Canberra against. Whether Australia ends up with a dedicated AI law or a reinforced version of the status quo will shape not just the technology sector, but the rights of every Australian who is increasingly governed, quietly, by code.
Sources: The Guardian. Greens digital rights spokesperson David Shoebridge.
