Australia established a government Office of AI within the Department of Prime Minister and Cabinet on 15 July 2026, the same day Prime Minister Anthony Albanese delivered a speech titled "AI in Australia's interests" (pm.gov.au).

The Australian government described the office as a "world first" for a joined-up approach to AI across government, according to RNZ (RNZ). Reuters reported the office is intended to coordinate AI regulation and make Australia a more attractive place for AI investment by giving more clarity for project approvals (Reuters).

Albanese promised legislation "early next year" to set rules for large data centres — the warehouse-scale facilities that house the computing power behind AI — regarding power and water use, per RNZ. The proposed laws would require data centres to produce net energy, in addition to limiting water use (ThePrint). The Australian Government posted on its official Facebook page that it will introduce a "world-leading" artificial intelligence framework (Facebook).

The Office of AI builds on earlier institutional groundwork. The Australian government announced on 25 November 2025 that a new AI Safety Institute would become operational in early 2026 (industry.gov.au). The National AI Centre launched the AI.gov.au digital platform in May 2026 to help organisations understand and use AI (industry.gov.au). The Australian Senate Select Committee on Adopting Artificial Intelligence was established on 26 March 2024 (industry.gov.au).

Across the Tasman, the contrast is sharp. The New Zealand government has adopted what RNZ describes as a "lighthanded approach" to AI while also aiming to rapidly spread its use across public agencies and businesses. But efforts to scrutinise what advice the government has actually received on AI's costs and benefits have met delays under the Official Information Act — the law that gives the public the right to request government information, with a statutory deadline of 20 working days for a response.

RNZ asked Finance Minister Nicola Willis for advice she had received about the costs or benefits of AI use in government. Willis transferred the request to Public Service and Digitising Minister Paul Goldsmith after 15 working days. Goldsmith's office waited 19 working days before advising RNZ the response was delayed by over another month, to 24 August.

Dr Andrew Lensen of Victoria University, the AI expert quoted in RNZ's reporting, described Australia's move as a "kick up the bum for New Zealand leaders." Lensen gave his comments to RNZ's Midday Report on a Thursday.

The broader context here is one of diverging institutional architecture. Australia has now placed AI coordination at the centre of its executive government, backed by a safety institute, a select committee process, and a dedicated digital platform. New Zealand's approach has been explicitly lighthanded, relying on diffusion through existing agencies and businesses rather than a central coordinating body. Whether that posture is sustainable in the face of a trans-Tasman neighbour building dedicated machinery of state is a question opposition MPs and policy analysts are likely to press.

The OIA delay is also worth noting for what it reveals about the current state of ministerial readiness on AI. A request moved between portfolios, bounced up against the 20-working-day statutory deadline, and was then pushed out by more than a month. For a government simultaneously promising rapid AI adoption across the public service, the difficulty producing any advice on costs and benefits within statutory timeframes is an instructive signal about how much substantive work has actually been done.

Goldsmith's office has not yet released the requested advice. The new deadline of 24 August, if met, would come more than two months after the original request was lodged.