Australia Goes Again with “World-Leading” AI Strategy? | American Enterprise Institute
Last week, Australian Prime Minister Anthony Albanese announced the country’s allegedly “world-leading” National AI Plan and associated measures to position AI as a strategic growth industry. This includes targeted public funding, infrastructure standards, and a new central governance structure: the Office of AI under supervision of the prime minister and cabinet. The framework is intended to attract AI investment. Furthermore, it will impose guardrails on data centers and public‑sector use of AI. Copyright enforcement for Australian content used to train foundation models will also be addressed.
The plan takes a “strategy-plus-standards” and largely technology‑neutral approach to governing aspects of AI applications within Australia’s boundaries. This contrasts with the EU AI Act’s prescriptive, risk-based regulation with extraterritorial obligations and formal market-access conditions. For providers serving both markets, governance calibrated to the EU AI Act will generally overshoot Australian requirements, with Australia adding some distinctive infrastructure sustainability and creator protection elements instead of creating a full AI‑Act analogue.
The strategy pictures Australia becoming a regional hub for sustainable AI infrastructure. It explicitly links industry policy to infrastructure and export‑oriented AI services. Central is the consolidation of more than A$460 million in existing AI-related government funding into a single industrial strategy aimed at growing a “world-class AI ecosystem,” intended to capture economic opportunities arising from the implementation of AI applications. This includes over A$360 million in research grants, A$47 million for the Next Generation Graduates Program, and A$40 million to expand the National Artificial Intelligence Centre and industry-adoption initiatives. The funding augments broader government-funded technology and industrial policy instruments, such as an additional A$1 billion under the National Reconstruction Fund for “critical technologies” and substantial AI‑related research and development tax incentive claims. Foreign investment is courted and will be necessary for development of high-cost infrastructure such as data centers.
A key strand of the plan is a national framework for AI data centers, with principles and eventual legislation to govern sustainability and resource use. The centers will be required to be energy-neutral, generating as much power as they consume, and to meet stringent water-efficiency expectations (important in a hot dry environment where water use is highly contestable). Office for AI-developed national data center principles will also set expectations on other factors such as job training and community sustainability, attempting to reconcile local resistance to energy‑intensive data centers with the AI policy economic growth imperative.
A strong stance on copyright and training data is embedded: Albanese declared that creators of books, music, art, and news “should retain control of the price and value of their work” when used to train AI, and that “anything less is theft.” The forthcoming national AI standards will include protections for creative professionals whose works are used to train AI models, though how this will be achieved is yet to be detailed. This connection between AI industrial policy and cultural‑sector bargaining power poses creator compensation as part of AI’s social license rather than as a purely IP‑law question – echoing Australia’s News Media Bargaining Code imposing contractual obligations on social media firms to compensate traditional media to support public interest journalism. How this affects existing copyright, contracts, and collective‑bargaining frameworks, however, remains opaque.
Positioning the new Office of AI under direct prime ministerial control ensures co-ordination across ministries and government control of standards development. This signals a shift toward more centralized governance without (yet) relying on EU-style legislative intervention. Unlike the EU, and more like the US approach, coordination of standards and policy will be centrally controlled, but supervision remains with existing regulators using their general powers rather than introducing any special ex ante AI regulatory powers.
Australia’s AI stance thus shifts from a largely “existing‑law-plus-voluntary-principles” approach toward a more coherent AI industry strategy with central coordination and public co‑investment, and with new hard‑law obligations on infrastructure and foreign AI companies utilizing Australian creative content. On the one hand, Australia seeks to attract foreign AI and data‑center capital and build a domestic export‑oriented AI ecosystem; on the other, it promises tighter constraints around environmental externalities and uncompensated use of cultural works.
But is Australia’s AI plan really world leading? It certainly differs from the EU approach to AI governance, but copyright issues aside, it is not clear that it differs from what is occurring in the United States: Centralized funding and control of industrial policy direction are occurring at state level as AI infrastructure and applications are deployed. Arguably, given the scale of the Australian economy, the incentives to protect that economy from competitive rivals and to grow it where possible to take advantage of what new technologies offer position it remarkably like, say, Louisiana or Arkansas in its AI policy approach.
