APS AI Tracker

Who’s who

No one desk runs Australian Government AI policy. The rules for AI in the economy sit in one department, and the rules for the government’s own use of AI in another. The delivery machinery sits in a third, and since July 2026 a new office in the Prime Minister’s own department sits across the lot. This page is a guide to who holds which piece, how the cast has grown, and where the transparency statements this site tracks fit in.

Unlike the rest of this site, which is generated from the daily scrape, this page is written and updated by hand. Every claim links to a government source, or to news coverage for announcements. Last reviewed 15 July 2026.

Two tracks, one new office

Most of the apparent sprawl resolves into two tracks. One sets policy for AI in the economy: what the businesses building and deploying AI should do. That work lives in the Department of Industry, Science and Resources. It runs from the AI Ethics Principles of 2019, through the safe-and-responsible-AI consultations of 2023–24, to the National AI Plan of December 2025.

The other track governs the government’s own use of AI: what the public service may do with these tools, and what it must tell the public. Three bodies in the Finance portfolio share this work, all answering to the Minister for Finance and the Public Service, Senator Katy Gallagher. The Digital Transformation Agency writes the rules, the Department of Finance builds the shared tools, and the Australian Public Service Commission looks after the workforce. The transparency statements this site tracks belong to this second track: they are required by a DTA policy.

The Office of AI announced on 15 July 2026 is a bid to coordinate both tracks from the Department of the Prime Minister and Cabinet. Its ambit also takes in the data-centre, copyright, education and national-security threads around them.

The cast

Digital Transformation Agency (DTA)

The DTA writes the rules for AI inside government, and it is the reason every statement on this site exists. An executive agency in the Finance portfolio, it descends from the Digital Transformation Office of 2015, renamed in 2016. With the industry department it published the first whole-of-government interim guidance on generative AI in July 2023, and it jointly led the AI in Government Taskforce (eighteen secondees from eleven agencies) from September 2023 to June 2024. It also ran the whole-of-government Microsoft 365 Copilot trial over the first half of 2024, across more than 60 agencies and around 7,600 staff.

Its central instrument is the Policy for the responsible use of AI in government, announced in August 2024 and effective that September, with a substantially strengthened v2.0 from 15 December 2025. Under it sits the Standard for AI transparency statements, which required agencies to publish a statement by 28 February 2025. Those statements are the documents whose every change this site records. The DTA also publishes a technical standard for government AI systems (July 2025) and an AI impact assessment tool with procurement guidance (December 2025); since March 2026 it keeps a central register of the transparency statements themselves. An AI Review Committee to oversee high-risk use cases across government was slated to be detailed by mid-2026. The CEO is Chris Fechner.

Department of Finance: AIDE and GovAI

Finance is the delivery arm. It led the AI Plan for the Australian Public Service, launched by Senator Katy Gallagher on 12 November 2025 around three pillars: trust, people, tools. Its coordination function is AI Delivery and Enablement (AIDE), a central multidisciplinary team stood up in late 2025. AIDE leads delivery of the plan and runs the shared platforms; it also reports on AI adoption across the service. Finance operates GovAI, the APS-only AI platform, piloted from 2024 and opened to the whole service on 31 July 2025. GovAI’s generative assistant, GovAI Chat, entered an alpha trial in April 2026, offering commercial models such as ChatGPT and Claude inside a government-managed environment.

The plan also required every non-corporate agency to appoint an SES-level Chief AI Officer by July 2026. AIDE convenes these officers, whose role is distinct from the accountable officials the DTA’s policy already required. By the deadline, 104 of the 106 agencies obliged to appoint one had done so.

Australian Public Service Commission (APSC)

The APSC owns the workforce piece. Under the APS AI Plan’s people pillar it leads foundational AI training, delivered through the APS Academy and completed by more than 100,000 public servants by July 2026. It also sets the standards for consulting staff on AI-related workplace change and hosts an AI Workforce Planning Taskforce; in April 2026 it published principles for AI in recruitment. Despite a common assumption, the 2023–24 AI in Government Taskforce was jointly led by the DTA and the industry department,not by the APSC.

Department of Industry, Science and Resources (DISR)

DISR sets AI policy for the economy at large, and it has housed the long debate over whether Australia should regulate AI directly. It published the AI Ethics Principles in 2019, developed with CSIRO’s Data61, and the AI Action Plan in 2021. Under Minister Ed Husic it ran the safe and responsible AI consultation of 2023, which drew more than 500 submissions. A January 2024 interim response promised guardrails for high-risk AI; an AI Expert Group followed, then a mandatory-guardrails proposals paper in September 2024.

The direction changed with the ministry. Husic was dropped in the May 2025 reshuffle; Senator Tim Ayres became Minister for Industry and Innovation and Minister for Science, with Dr Andrew Charlton as Assistant Minister for Science, Technology and the Digital Economy. TheirNational AI Plan of 2 December 2025 chose existing legal and regulatory frameworks over a dedicated AI Act, and the mandatory-guardrails proposal did not proceed. DISR also carries Australia’s international AI-safety commitments, from the Bletchley Declaration of 2023 onward.

National AI Centre (NAIC)

The NAIC is the industry-facing centre: the voluntary-guidance counterpart to the DTA’s mandatory rules for agencies. It was launched on 14 December 2021 inside CSIRO’s Data61 under the Morrison government’s AI Action Plan, with Stela Solar as inaugural director. It convened the Responsible AI Network in 2023 before a 2024–25 Budget decision moved it out of CSIRO and into DISR. Its Voluntary AI Safety Standard of September 2024 set out ten voluntary guardrails, since folded into the leaner Guidance for AI Adoption of October 2025. Led by Lee Hickin since March 2025, it is a key delivery body for the National AI Plan and runs the government’s public AI portal, ai.gov.au.

Australian AI Safety Institute

The institute is the newest of the DISR bodies, announced on 25 November 2025 with $29.9 million and operational from early 2026. It tests advanced AI systems, assesses their risks and harms, and maps regulatory gaps. That work is advisory; the institute has no enforcement powers. It joined the international network of AI safety institutes.

Office of AI (Prime Minister and Cabinet)

The Office of AI is the newcomer. In a speech on 15 July 2026, Prime Minister Anthony Albanese announced the office within his own department and declared the end of the government’s “issue-by-issue” approach. Australia, he said, would be the first country to bring AI policy into “a single, national framework”. The office is to coordinate the design of Australian standards for the technology and to pull together the threads running through other ministers’ portfolios: data-centre approvals with the states, copyright, AI at work and in schools, defence and national security. It will work with Minister Ayres and Assistant Minister Charlton. At the time of writing, no head, budget or staffing had been announced, and how the office relates to AIDE, the DTA or the NAIC had not been spelled out. The opposition dismissed it as “an office in an office … inside his own office”.

Around them

A ring of other bodies has a hand in the policy without owning it. The Data and Digital Ministers Meeting, the Commonwealth–state forum, agreed a national framework for assuring AI in government in June 2024. Parliament has weighed in repeatedly. The Senate Select Committee on Adopting AI recommended a dedicated AI Act in November 2024; the government’s response arrived in April 2026, after the National AI Plan had gone the other way. The audit committee’s Proceed with Caution report of March 2025 called for standing parliamentary oversight of AI, and a Senate inquiry into AI and data centres opened in May 2026. The regulators publish AI guidance within their existing remits rather than holding policy carriage: the privacy regulator on AI products (2024), eSafety through industry codes covering generative AI and companion chatbots (2025), the signals directorate on secure use of AI (2024), and the national audit office through agency-by-agency AI governance audits. On copyright, the issue that shadows every data-centre deal, the Attorney-General’s Department convenes the Copyright and Artificial Intelligence Reference Group.

How it grew

A hand-picked chronology of the moments that built the current arrangement. Dates are the official publication or announcement dates in the linked sources.

Corrections

Machinery of government moves faster than hand-curated pages. If something here is out of date or wrong, please open an issue.