Briefings

Australian AI Governance Briefing: Week Ending 30 August 2026

National Cabinet agreed to develop consistent mandatory standards for data centre energy, water and land use and confirmed the Commonwealth intends to legislate national AI standards in early 2027, while APRA and ASIC issued a joint call for the financial sector to move from awareness of frontier AI risk to tested response plans.

4 stories

A substantial week, with movement on three fronts. National Cabinet met in Sydney on Wednesday 26 August and agreed that the Commonwealth will work with states and territories to develop consistent mandatory standards for data centre energy, water and land use, confirming an intention to legislate national AI standards in early 2027. The energy-source question was not settled: leaders agreed to further work, and Queensland and the Northern Territory pressed for flexibility to use coal and gas rather than a renewables-only requirement.

The day after, APRA and ASIC published a joint call for financial market entities to move from awareness of frontier AI risks to tested action, drawing on nine roundtables held in June and July with more than 600 attendees. The regulators released an accompanying information paper containing a preparedness checklist for boards and executives, and identified third-party concentration risk as a sector-wide concern.

On Tuesday 25 August, Australian Information Commissioner Elizabeth Tydd used a keynote to the Australian Law Librarians’ Association Conference in Melbourne to press government agencies on automated decision-making transparency, ahead of the new APP 1.7 to 1.9 obligations commencing on 10 December 2026. Late in the week, the Government launched the Buy Australian AI Partnership, connecting Australian AI companies with the major banks as enterprise buyers.

The week in review

National Cabinet agrees to mandatory data centre standards and confirms early-2027 AI legislation

National Cabinet met in Sydney on Wednesday 26 August 2026. On artificial intelligence, First Ministers agreed that large data centres bring material energy, water and land-use impacts that need to be managed, and backed the Commonwealth’s plan to introduce a nationally consistent regulatory framework setting minimum requirements for large data centres.

The communique records that the Commonwealth will work with state and territory governments to develop consistent mandatory standards for data centre energy, water and land use, and to support skills and training opportunities. It states that the Commonwealth intends to legislate the AI standards in early 2027, including conditions associated with delivering AI training, and that Commonwealth legislation will be designed to complement rather than duplicate state and territory planning and approval processes.

The energy-source question was left open. The Commonwealth had pushed for a standard requiring large data centres to be net generators of power, drawing on renewables. Leaders did not settle that point and agreed to undertake further work. Queensland Premier David Crisafulli pressed for the ability to power Queensland data centres with coal or gas, noting that the state owns its distribution, transmission and generation assets; Northern Territory Chief Minister Lia Finocchiaro argued for flexibility given the Territory’s gas resources. Energy Minister Chris Bowen subsequently indicated that renewables could be backed by gas firming, and that state-owned fossil-fuel generators may be able to apply to provide primary electricity to facilities.

The load context comes from AEMO’s 2026 Electricity Statement of Opportunities, which reports that around 165 data centres currently consume approximately 5TWh of electricity a year, about 3% of National Electricity Market operational consumption, and forecasts that rising to 34TWh, or 13% of NEM operational consumption, by 2035-36.

The agreement follows the Prime Minister’s 15 July 2026 announcement of legislated Australian Standards for AI and an Office of AI within the Department of the Prime Minister and Cabinet, and the voluntary Expectations of data centres and AI infrastructure developers released on 23 March 2026. Coverage thresholds, the definition of a large data centre, the substance of the energy, water and land-use requirements, and commencement all remain to be developed with the states and territories.

Primary sources: Prime Minister of Australia — Meeting of National Cabinet, 26 August 2026 | ABC News — Gun licences restricted to citizens, data centre carve-out

APRA and ASIC tell financial entities to move from frontier AI awareness to tested action

APRA and ASIC published a joint release on 27 August 2026, ASIC media release 26-201MR, urging financial market entities to move from gaining awareness of risks linked to frontier AI to taking decisive action. It builds on nine roundtables the two regulators hosted in June and July 2026 involving more than 600 attendees from across the financial system. The Australian Signals Directorate supported the sessions, which also included the Reserve Bank of Australia, Treasury and the ACCC — participation the regulators described as signalling a whole-of-government response.

An accompanying information paper, Insights from the APRA-ASIC Industry Roundtables, sets out the themes and includes a preparedness checklist for boards and executives. Those themes were: getting the cyber fundamentals right, including identifying and managing critical assets and systems, timely patching, strong identity and access controls, attack surface reduction, backup integrity, tested response and recovery arrangements, and third-party risk management; settling key decisions such as risk appetite, escalation authority, recovery priorities and communication strategies at board level before a crisis, on the basis that frontier AI compresses incident response timeframes; growing interest in defensive AI for threat intelligence, vulnerability detection, code review and incident response, with an acknowledgement that capability remains limited; third-party concentration risk, where shared dependence on service providers could turn an isolated incident into broader financial-sector disruption; and industry-led collaboration and threat-intelligence sharing.

ASIC Commissioner Simone Constant said the urgency of the challenge could not be overstated, that threat actors are exploiting frontier AI models to identify and exploit vulnerabilities that previously may have taken a team of professionals months to find, and that boards and executives must move beyond awareness and ensure their organisations have well-tested response plans and understand where they are vulnerable. APRA Deputy Chair Therese McCarthy Hockey said this was the first time APRA and ASIC had created forums for rapid information-sharing across such a broad cross-section of the financial sector, and pointed to the willingness of more advanced entities to share practical approaches with less mature peers.

The release continues a line the two regulators have been running since APRA’s letter to industry on artificial intelligence in April 2026 and ASIC’s 8 May 2026 letter to AFS licensees and market participants, media release 26-092MR, which framed cyber resilience as a core licensing obligation rather than an IT issue.

Primary sources: ASIC — 26-201MR ASIC and APRA warn frontier AI awareness must turn to action | Finextra — ASIC and APRA warn frontier AI awareness must turn to action

Information Commissioner presses agencies on automated decision-making disclosure

Australian Information Commissioner Elizabeth Tydd delivered a keynote address to the Australian Law Librarians’ Association Conference in Melbourne, published by the OAIC on 25 August 2026 under the topic of information governance. The address argued that strong information governance, transparency and human-centred oversight must remain central as AI and automated decision-making are adopted more widely across government and legal institutions, and that AI should not undermine transparency or human-centred decision-making. It also raised AI-generated hallucinations in court filings and the need for safeguards and disclosure requirements.

Tydd returned to the findings of the OAIC’s report Automated decision-making and public reporting under the Freedom of Information Act, which followed a desktop review conducted in October 2025 of the websites of 23 government agencies authorised to use ADM. Of those 23 agencies, four — 17% — disclosed their use of ADM in their Information Publication Scheme material as operational information required to be published under the FOI Act. A further nine referred to ADM in their IPS disclosures without confirming that they used it, and those that did disclose were not clear about how they used it.

The address was delivered against the 10 December 2026 commencement of the new automated decision-making transparency obligations in APP 1.7 to 1.9, under which APP entities whose computer programs make, or do a thing substantially and directly related to making, decisions that significantly affect an individual’s rights or interests must set out in their privacy policies the kinds of personal information used and the kinds of decisions made. The OAIC ran an issues paper consultation on transparency in automated decision-making, opened on 18 May 2026, with guidance expected before commencement.

Primary sources: OAIC — Australian Law Librarians’ Association Conference | OAIC — Australian Information Commissioner highlights improved transparency and integrity for government agencies in automated decision-making | OAIC — Consultation on Guidance for Transparency in Automated Decision Making

Buy Australian AI Partnership launched with the major banks as founding enterprise partners

Stone & Chalk announced the Buy Australian AI Partnership on 27 August 2026, with the Government’s own announcement following on 28 August. The initiative was launched by Assistant Minister for Science, Technology and the Digital Economy Dr Andrew Charlton MP and is delivered by Stone & Chalk, with the National AI Centre as Principal Sponsor and ANZ, Commonwealth Bank of Australia, Cuscal, National Australia Bank and Westpac as Founding Enterprise Partners. The broader coalition includes Gradient Institute, the Customer Owned Banking Association, the Australian Banking Association, the Insurance Council of Australia, the AIIA, FinTech Australia, CEDA, Superteam Australia and MYOB.

The structure has two stages. A national cohort receives expert advice and enterprise engagement, and selected companies then complete an intensive eight-week accelerator covering AI assurance, regulatory requirements, procurement navigation and technology implementation. The first intake focuses on financial services, with the founding banks and Cuscal helping participants understand enterprise procurement, compliance and deployment requirements. Expressions of interest close on 24 September 2026, with the program running from early October 2026.

Stone & Chalk CEO Stela Solar said Australia already has the talent, the technology and the startups, and that Australian enterprises are now opening their doors to homegrown innovation. The Government positions the initiative as supporting the National AI Plan’s commitment to strengthen Australian AI capability and create stronger pathways between innovation and adoption, and as reducing reliance on overseas-owned technology.

Primary sources: Stone & Chalk — New Australian AI Partnership to connect homegrown AI companies with financial services buyers | ABC News — Australia in a ‘sliding doors’ moment to determine who reaps rewards of AI boom

Stories

National Cabinet backs mandatory national data centre standards, with legislation flagged for early 2027

National Cabinet met in Sydney on 26 August 2026 and agreed that the Commonwealth will work with state and territory governments to develop consistent mandatory standards for data centre energy, water and land use, backing a nationally consistent framework setting minimum requirements for large data centres. The communique states that the Commonwealth intends to legislate the AI standards in early 2027, including conditions associated with delivering AI training, and that the legislation will be designed to complement rather than duplicate state and territory planning and approval processes. The energy-source question was not resolved, with leaders agreeing to further work after Queensland and the Northern Territory pressed for the ability to use coal and gas. Coverage thresholds, the substance of the standards and commencement all remain to be developed.

Source: pm.gov.au

APRA and ASIC issue joint call for financial entities to act on frontier AI risk

APRA and ASIC published a joint release on 27 August 2026, ASIC media release 26-201MR, urging financial market entities to move from awareness of frontier AI risks to decisive action, together with an information paper drawn from nine roundtables held in June and July 2026 with more than 600 attendees. The paper includes a preparedness checklist for boards and executives and identifies third-party concentration risk, limited defensive AI capability, and the compression of incident response timeframes as key themes. ASIC Commissioner Simone Constant said boards and executives must move beyond awareness and ensure their organisations have well-tested response plans. The roundtables were supported by the Australian Signals Directorate and included the RBA, Treasury and the ACCC.

Source: asic.gov.au

Information Commissioner presses agencies on automated decision-making transparency ahead of December

Australian Information Commissioner Elizabeth Tydd delivered a keynote to the Australian Law Librarians’ Association Conference in Melbourne, published by the OAIC on 25 August 2026, arguing that transparency, accountability and strong information governance must remain central as AI and automated decision-making are adopted across government and legal institutions. She returned to the OAIC’s finding that of 23 government agencies authorised to use ADM, only four — 17% — disclosed that use in their Information Publication Scheme material as required under the FOI Act. The address also raised AI-generated hallucinations in court filings and the need for safeguards and disclosure. New ADM transparency obligations in APP 1.7 to 1.9 commence on 10 December 2026.

Source: oaic.gov.au

Government launches Buy Australian AI Partnership with the major banks

The Buy Australian AI Partnership was announced by Stone & Chalk on 27 August 2026 and launched by Assistant Minister Dr Andrew Charlton MP, connecting Australian AI companies with enterprise buyers in banking, insurance and superannuation. Stone & Chalk delivers the program with the National AI Centre as Principal Sponsor and ANZ, Commonwealth Bank, Cuscal, NAB and Westpac as Founding Enterprise Partners. Selected companies complete an eight-week accelerator covering AI assurance, regulatory requirements, procurement navigation and technology implementation, with the first intake focused on financial services. Expressions of interest close on 24 September 2026 and the program runs from early October 2026.

Source: stoneandchalk.com.au


This briefing was researched and written with AI assistance.

Stay across Australian AI governance

Get the briefing delivered to your inbox every week. No spam, unsubscribe any time.