The Australian Public Service AI Plan 2025: A legal and commercial roadmap

Pillar One: Trust – Transparency, Ethics, and Governance

Pillar One focuses on establishing a strong regulatory and ethical environment to maintain public confidence and manage risk. Such an environment is critical to establishing the APS’s “license to operate” for AI use.

Key Initiatives and Legal Implications

Policy and guidance updates

The Policy for Responsible Use of AI in Government is being updated by the DTA to strengthen accountability, enhance risk identification, and mandate the use of the AI Impact Assessment tool for in-scope use cases. Agencies will also be required to develop and communicate a strategic position on AI adoption.

Enhanced Oversight

The Government (led by the DTA) will establish an AI Review Committee, comprised of APS experts, to ensure consistent, responsible deployment of AI across the APS and enhance whole-of-government oversight. The committee will provide non-binding advice on sensitive and high-risk AI deployments, ensuring cross-disciplinary scrutiny.

Contractors and External Providers

New clauses will be added to the Commonwealth Contracting Suite and Clausebank to:

• Require suppliers on the Management Advisory Services and People Panels to advise of any planned use of AI when quoting for services.
• Clarify that consultants and external contractors remain responsible for the services they deliver, regardless of whether generative AI was used.
• Ensure transparency and accountability in external providers’ use of generative AI technologies.

Pillar Two: People – Capability Building and Engagement

The second pillar of the APS AI Plan addresses the workforce transformation required to use AI safely, responsibly and effectively. The focus is on skills, leadership, and change management at a whole-of-APS level.

Key Initiatives

Foundational Learning

Foundational AI literacy training offering will be mandated for all staff through the updated AI in Government Policy, led by the APSC. This aims to provide public servants with the capability to use AI safely, responsibly, and effectively.

Leadership and Adoption

Agencies will appoint Chief AI Officers (CAIOs) – senior executive leaders tasked with accelerating consistent capability development, driving cultural change, and overseeing adoption and innovation within their agencies.

Workforce Consultation

The APSC will issue a Circular setting out standards for consultation on AI-related workplace changes, aligning with existing APS Enterprise Agreements and engagement mechanisms. This ensures meaningful staff and union input on changes.

AI Delivery and Enablement (AIDE)

A central multidisciplinary team will be established to accelerate uptake, help tackle common adoption barriers and share use cases for re-use across government. AIDE will be designed to complement rather than replace existing structures and processes across whole-of-government.

Pillar Two: People – Capability Building and Engagement

Pillar Three focuses on providing the technical foundation for AI adoption: secure, fit-for-purpose AI technologies. This will be achieved through the provision of secure AI tools for sensitive activities, while public web-based AI can be used for low-risk activities.

Key Initiatives

Centralised Infrastructure

GovAI will be leveraged to provide a secure, central, Australian-based platform for developing customized AI solutions, preventing vendor lock-in, and enabling access to diverse AI models (including an onshore instance of OpenAI’s GPT models). Custom AI solutions will be available for specialist workflows and use cases.

Universal, Secure Access

The government will provide GovAI Chat – a universally accessible, secure, purpose-built AI assistant – to everyone in the APS. This initiative aims to leverage government data while complying with legal and cyber security frameworks.

Guidance on Public Tools

Clear guidance will be developed on using public generative AI tools (such as ChatGPT, Claude, and Gemini) up to OFFICIAL level information, allowing staff to access general-purpose tools for low-risk activities.

Procurement Efficiency

A central register of generative AI assessments will be created on GovAI to share completed security, FOCI, and impact assessments, allowing agencies to reuse prior evaluations and streamline procurement.

Key Milestones

Conclusion: Legal and commercial implications

The APS AI Plan 2025 is a significant development in the governance arrangements and workforce management in the Commonwealth for the use of generative AI. The Plan lays the foundation for a modern, responsive public service that captures the opportunity of AI while ensuring ethical and responsible guardrails remain fit-for-purpose.

The development is also positive for those engaging with the Commonwealth.

The development of AI tools for government is one example of the mutually beneficial relationship between Government and industry, and how such collaborative partnerships can improve the quality and efficiency of the delivery of government services. As AI tools proliferate and improve, the AI maturity of the Commonwealth will assist industry to better understand how it can responsibly deploy the use of AI tools in the provision of services to government. Notably, clarity on contractor responsibility and the new procurement transparency requirements will help both Commonwealth agencies (on the one hand), and contractors and services providers who engage with the Commonwealth (on the other), to assess and manage legal risk throughout the procurement lifecycle.