Ross Stitt reports on how the Australian Government is facing up to the AI challenge of how to secure productivity benefits without the electoral cost of increased unemployment and loss of sovereignty

Australian Prime Minister Anthony Albanese is a huge fan of his 2025 ban on social media for Australians under the age of 16. It was a world first that is now being implemented in many other countries.

Fresh from his self-proclaimed success in tackling the challenges of social media, despite reports that most children get around it, Albanese has now shifted his attention to an even greater threat – artificial intelligence (AI).

In a significant speech delivered this month at the University of Sydney, the PM declared that ‘this is our time to decide what AI looks like here in Australia’ and to ‘make AI work in Australia’s national interests’.

That’s a big call.

But Albanese’s an astute politician. He recognises the rising anxiety in the community about both the speed and the scope of the emerging AI revolution. Ignoring it is not a politically viable option.

In May, the Office of the Australian Information Commissioner (OAIC) released the Australian Community Attitudes to Privacy Survey 2026. It illustrates the grave misgivings many Australians have about AI.      

According to the survey, just 4% of Australians trust AI companies and 96% say conditions should be placed on the use of AI.

The survey addresses two specific concerns – the role of AI in decision-making by organisations and the privacy implications of AI.

AI is already being used in the private sector and government to review job applications and procurement contracts. And it has significant potential for assessing entitlements to social welfare benefits.   

But the idea of computers deciding the fate of humans is unnerving for many. 

81% of respondents to the OAIC’s survey want ‘the right to have a human review’ an AI decision and 68% want AI decision-making tested for ‘bias and discrimination’.    

On the privacy front, 76% of Australians want limits on the ability of AI providers to retain personal information and 76% do not want their personal information used ‘to train external AI systems’.

The government has already moved to address many of these concerns. On 20 July six government ministers issued a joint statement outlining a raft of new ‘AI consumer safety priorities’.  These include imposing a duty of care on AI companies, reforming privacy rules, introducing new consumer protections, and regulating ‘the use of automated decision-making within federal agencies’.

There are two additional concerns that dominate the AI discourse in Australia – the threat of widespread unemployment as AI replaces workers, and the threat of environmental damage from the water and energy demands of data centres.

Earlier this month the Department of Employment and Workplace Relations published a report examining whether AI is affecting employment. Its conclusion is that ‘there is no evidence to date of broad AI-driven labour-market upheaval in Australia’.         

That’s clearly not the view of the Australian Council of Trade Unions. Two weeks ago, ACTU secretary Sally McManus claimed that ‘too many big employers have introduced AI into workplaces without transparency and used it to justify slashing jobs’. In her view ‘Australians don’t trust AI because of the lack of clear benefit for working people’.

The challenge for any government, particularly a Labor one, is obvious – how to secure the productivity benefits of AI without incurring the electoral cost of increased unemployment.

Prime Minister Albanese argued in his University of Sydney speech that ‘we should not treat AI as a threat to good jobs - we must use it as an instrument to help create them’. Unsurprisingly, his speech said little about exactly how that will be achieved.

Albanese was much more specific on how his government will manage the environmental threat from data centres. There will be a national regulatory framework that establishes clear rules. Crucially, data centres will have ‘to be net-generators, not net-users’ of energy. They will need ‘to build new renewable generation’.

The key requirement identified by Albanese is that ‘data centres do not increase power prices for Australians’. He doesn’t want unhappy voters.

The obvious flaw in Albanese’s proposal is that the development and construction of a wave of new data centres is already well underway. The delivery of new renewable electricity generation sufficient to service those data centres is not.

As recently acknowledged by OpenAI CEO Sam Altman, Australia has the potential to be a significant player in the data centre market. The Australian PM recognises the economic benefits that would bring. By establishing a national regulatory framework, he hopes to ‘enhance our appeal to international investors by delivering greater clarity and speed for approvals’.

But it will not be easy balancing the objectives of international investors with the concerns many Australians have about data centres.

Perhaps the most intriguing element of Albanese’s speech was his view on the role AI must play in building Australian national sovereignty and economic resilience. He says it should not be a case of Australia ‘just adopting or accommodating AI’, but rather ‘designing it, making it, building the capability right here’. He argues that ‘if we are always dependent on someone else, somewhere else, we will always be vulnerable’.

This touches on one the great unknowns of the AI revolution. Will the US AI trailblazers continue to dominate the field and, if so, in what form will they share their advanced AI models? Can a country like Australia achieve any kind of AI sovereignty or will it remain dependent on, and therefore vulnerable to, US entities?

China is seen as not far behind the US in the development of AI. And to date it has appeared more willing to promote open-source AI than the US tech giants.

Earlier this month President Xi said, ‘AI development should not be a solo performance by a single country, but a symphony of international cooperation’. To that end China has formed a coalition of 29 countries called the World Artificial Intelligence Cooperation Organisation.

Australia is unlikely to join WAICO any time soon. But if the Australian government wants to ensure national sovereignty and economic resilience, the advantages of open-source AI are obvious.  

On a more ominous note, this month saw an OpenAI experimental model ‘break out’ of a supposedly isolated environment and hack into the system of an unrelated entity, US startup Hugging Face. This hints at a dystopian future in which AI operates independently of its human ‘masters’.   

In such a future, AI will be beyond the control of governments

*Ross Stitt is a freelance writer with a PhD in political science. He is a New Zealander based in Sydney. His articles are part of our 'Understanding Australia' series.