AI Policy overhauled with new Impact assessment tool and Procurement guidance

The AI Plan for the Australian Public Service 2025 sets out a roadmap for the responsible adoption of artificial intelligence by the Australian Government. Within it, the DTA was assigned several deliverables to assist the public service in its efforts to experiment and scale this technology to improve service delivery.

These initiatives are:

Updating AI in Government Policy

The DTA has updated its Policy for responsible use of AI in government to strengthen public trust in government.

“We’ve ensured that this new version of our policy provides clarity, enhances accountability, and ensures effective risk identification and management,” explains Lucy Poole, Deputy Chief Executive Officer and General Manager of Strategy, Planning and Performance, DTA.

“This builds on our original Policy which laid the foundation of accountability measures, transparency requirements, and guidance for foundational training.”

As AI adoption across the APS accelerates and the technology landscape evolves, the DTA update includes new requirements for agencies to develop a strategic position on AI adoption and to communicate this position to staff. This will support agencies to better engage with and realise the benefits of AI.

Accountability requirements have been strengthened so each use case has a clearly assigned accountable officer and is recorded in an internal register. 

The update also builds trust through mandating an AI impact assessment for use cases. Impact assessments allow for appropriate governance and risk management actions, including for higher-risk use cases.

The updated Policy for responsible use of AI in government comes into effect Monday 15 December 2025.

New AI impact assessment tool

Our AI impact assessment tool, and supporting guidance, is designed to help government teams identify, assess and manage AI use case impacts and risks against Australia's AI Ethics Principles.

“Understanding and managing AI use case impacts and risks is critical for effective AI governance,” stresses Ms Poole. “Our approach to AI is designed to foster innovation by giving agencies the space to explore and experiment responsibly.”

“It creates opportunities to test and learn within clear guardrails that uphold APS values and maintain public trust. This environment ensures that experimentation is not only encouraged but guided by strong principles of safety, accountability, and integrity.”

The impact assessment tool and its supporting guidance are intended to complement and strengthen – not duplicate – existing frameworks, legislation and practices that relate to government AI use. It does this by focusing on AI-specific impacts and risks that existing approaches may not fully address. It does not replace a comprehensive risk management plan, which captures all risks, treatments and ongoing monitoring measures. 

Agencies may also use this tool to support other AI-related activities, such as procurements where suppliers use AI to provide goods and services. 

Guidance on AI procurement in government

The Guidance on AI procurement in government helps government buyers confidently procure AI products and services, including those that involve the use of AI. It provides clear, step-by-step advice on procuring AI in line with the Digital Sourcing Lifecycle: Plan, Source, and Manage.

It expands on each stage of the procurement process by highlighting AI-specific risks and considerations, complementing the DTA’s broader suite of frameworks. This includes the Policy for the responsible use of AI in government, the AI technical standard and the AI impact assessment tool.

Features of the guidance include steps to:

• Define clear objectives and a develop a business case. Objectives should focus on outcomes rather than prescribing a single technical solution.
• Review relevant policies, laws and frameworks. Agencies should consider obligations across risks such as privacy, security, and ethics that shape how AI can be procured and used, to strengthen assurance.
• Engage multidisciplinary teams. Perspectives from technical, legal, procurement and governance experts, as well as stakeholders and end-users, can better support agencies through procurement processes.
• Assess the readiness of data and infrastructure.  Agencies should confirm data is sufficient, high quality and appropriate, and systems can support the proposed AI solution.

When used alongside the AI procurement checklist, the guidance gives agencies a structured way to approach complex procurement decisions, allowing them to move through each stage with clarity and confidence.

“Agencies need confidence and clarity at every stage of AI adoption, and procurement is a critical turning point where planning meets implementation,” said Ms Poole. “This guidance and checklist equip them with the practical tools required to buy responsibly and deliver outcomes Australians can trust.”

Up next

The DTA is continuing to finalise the terms of reference for the AI Review Committee. The Committee will advise on high-risk AI use cases across the APS, ensuring responsible, consistent, and safe deployment. It will review sensitive cases, offer guidance on emerging risks, and support continuous improvement in government AI practices.

More details will be released in Q1 of 2026.

We will be releasing deep dive articles into each of these projects. Subscribe to our mailing list and LinkedIn page to be notified when our articles come out.

In the meantime, explore all of our latest AI resources below:

• Policy for responsible use of AI in government

• AI impact assessment tool and supporting guidance

• Guidance on AI procurement in government and check list

• AI Plan for the Australian Public Service 2025