WA Government AI Policy and AI Assurance Framework

What it is

The Government of Western Australia’s Artificial Intelligence Policy and AI Assurance Framework support the safe, secure and reliable use of AI in the WA public sector. They are published by the Office of Digital Government.

The AI Policy sets out the principles behind WA’s approach to using AI responsibly. It encourages innovation while making sure risks are managed, and requires each government entity to nominate an AI Accountable Officer.

The AI Assurance Framework is a self-assessment that helps entities evaluate an AI project’s implications, risks and controls in line with the policy’s principles and the national framework for the assurance of AI in government.

Who it applies to

WA government entities. The self-assessment must be used for any system or project driven by AI or automated decision-making, and applied at every phase, whether the AI system is procured, built or otherwise sourced or adapted. It is completed by the lead of the AI project, or the person implementing a new AI tool.

What it requires

An AI Accountable Officer. Each entity nominates an executive responsible for overseeing how AI is used. The officer might come from information technology, information management or data science and analytics. They make sure the entity’s AI use aligns with the policy and approve its self-assessments. Nominations were due by close of business on 19 September 2025.

A self-assessment for every AI project. The project lead completes the self-assessment through the online form. It guides them through describing the project and its benefits, assessing privacy, security, transparency, explainability and contestability risks and outlining risk controls. A downloadable template helps with drafting. The form assigns an outcome and sends the assessment to the AI Accountable Officer for approval. Both receive a PDF copy. The form can be reopened and resubmitted at later stages of the lifecycle.

Board review for higher-risk or larger projects. A completed self-assessment must go to the WA AI Advisory Board if any of these apply.

• Residual risk, after mitigations, is mid-range or higher.
• The project is funded through the Digital Capability Fund.
• The total cost exceeds A$5 million.

Risk outcomes and review

Low-risk projects need no further action once the self-assessment is approved. Medium and high risk projects go to the AI Advisory Board for review. The Office of Digital Government contacts the project and may ask for supporting documents, such as a detailed risk assessment, high-level solution architecture, functional and non-functional requirements, the business case, a privacy impact assessment and material on governance, change management and training.

Projects can continue while under review. The Board gives recommendations on managing risk and other considerations in a letter to the entity’s AI Accountable Officer and chief executive, and may ask for updates as the project progresses. Its advice is non-binding, but entities are recommended to consider it fully.

What it means for suppliers

The policy binds WA government entities, not their suppliers. A supplier of an AI system or tool to a WA entity should still expect to help with the self-assessment. The project lead has to describe the system and its benefits, assess its privacy, security, transparency, explainability and contestability risks and set out controls. If the project goes to the AI Advisory Board, the Office of Digital Government may ask for solution architecture, requirements and a privacy impact assessment. A supplier that can describe its system, its data handling and its human oversight clearly makes that work faster.

Key dates

The policy and framework were first published on 5 March 2024. AI Accountable Officers had to be nominated by 19 September 2025. The page was last updated on 25 August 2026. The Board’s meeting cut-off dates for submissions are published by the Office of Digital Government.

How it relates to other frameworks

The WA self-assessment aligns with the national framework for the assurance of AI in government, which is built on Australia’s AI Ethics Principles. Its structure resembles the NSW AI Assessment Framework, with an online assessment, a risk outcome and expert review for higher-risk and higher-cost projects. NSW’s review is triggered by high or critical risk, and WA’s by mid-range residual risk or above. Queensland agencies use the FAIRA instead. Like the Commonwealth DTA policy, WA requires a named accountable executive, which makes ownership the first thing to settle before any self-assessment is started.