New AI roadmap to modernise Australian public service

The Australian Government has released its AI Plan for the Australian Public Service 2025, outlining a framework to expand the safe and responsible use of artificial intelligence across federal agencies. The plan aims to improve service delivery, policy outcomes and workforce productivity through broad access to secure AI tools and comprehensive staff training.

Every public servant will receive foundational training in the safe use of generative AI, supported by clear guidance and leadership from newly appointed Chief AI Officers in each agency. These senior executives will oversee implementation, drive adoption and report on progress.

"The Plan sets a proactive stance to increase AI adoption. It ensures the Australian Government keeps pace with community expectations as well as international peers," said Senator Katy Gallagher, Minister for Finance.

Building trust

The plan is structured around three pillars – Trust, People and Tools. The Trust pillar focuses on transparency, ethics and governance. It includes updates to the government's policy for the responsible use of AI, requiring agencies to maintain internal registers of AI applications and to appoint accountable officers for each use case.

A new AI Review Committee will enhance oversight of high-risk applications, with members drawn from across the public service and external regulators. The committee will provide non-binding recommendations on complex or sensitive AI deployments and monitor emerging risks.

The government will also strengthen accountability among suppliers. Service providers will be required to disclose any use of AI when delivering services, and contract clauses will clarify that responsibility for outcomes remains with the supplier, regardless of whether AI tools are used.

Workforce capability

The People pillar addresses the skills and cultural changes required for adoption. All staff will complete mandatory foundational AI literacy training, supported by resources through the APS Academy. Senior leaders will participate in dedicated learning programs to guide responsible AI use within their departments.

Consultation with employees and unions will be a core component of implementation. The Australian Public Service Commission will issue a circular outlining standards for engagement on AI-related workplace changes, ensuring that staff and their representatives can provide input on how the technology affects their work.

A new unit, AI Delivery and Enablement (AIDE), will coordinate adoption across government. The team will identify and address barriers, share lessons from early adopters and develop practical frameworks to accelerate implementation while maintaining safeguards.

Technical foundations

The Tools pillar establishes the digital infrastructure needed for consistent and secure AI use. The GovAI platform will provide agencies with a centralised, Australian-based service for developing and hosting AI applications. It will include model brokerage, shared training resources and a library of approved use cases, helping to prevent duplication and reduce costs.

A companion service, GovAI Chat, will give all public servants access to a secure generative AI tool designed for government use. The system will allow staff to use AI for routine tasks while protecting sensitive information and ensuring compliance with privacy and data protection requirements.

Guidance will also be issued to help public servants safely use public generative AI platforms such as ChatGPT, Claude and Gemini for unclassified information. The Department of Home Affairs and the Digital Transformation Agency will oversee the related protocols and training.

Further initiatives include a central register of AI risk assessments, updates to AI procurement processes, and a new whole-of-government cloud policy to support responsible data management and secure computing.

Ethical framework

The accompanying Statement of Intent on AI in the Australian Public Service reaffirms that AI adoption must prioritise public benefit and equity. It commits the government to using AI to improve services, efficiency and productivity while ensuring its use is fair, transparent and aligned with community expectations.

Public confidence, the statement says, depends on ethical and legal safeguards that keep pace with rapid technological change. It emphasises that AI should be used to benefit people rather than for its own sake.

"The Government will always put people first and ensure AI adoption doesn't exacerbate inequality, bias, or unfair treatment," said Gallagher.