WA Government AI Governance Toolkit: What Public Sector Leaders Need to Know

The WA Government published its AI Governance Toolkit in August 2026, establishing a structured governance framework for artificial intelligence use across the Western Australian public sector. The toolkit applies to any system or project that is driven by AI or automated decision-making, and it covers the full lifecycle of those systems, from initial proposal through procurement or development to operation and review. For WA public sector entities that have adopted AI tools, are planning to adopt them, or are procuring systems from vendors that include AI capabilities, the toolkit creates immediate obligations that need to be understood and acted on.

The first and most concrete requirement is the nomination of an AI Accountable Officer. Every WA public sector entity must designate an executive who is responsible for aligning AI use with the WA AI Policy, approving self-assessments for AI systems, managing AI-related risks, and maintaining an AI tools reference list and use case register. This role requires expertise in information technology, data management, or analytics. The Accountable Officer also serves as the entity's liaison with the Office of Digital Government. Nominating this officer is not optional, and the role carries genuine accountability: it is an executive function, not a delegation to a technology team.

The self-assessment requirement is the operational core of the toolkit. Every AI system or project must complete a self-assessment at all phases of the project, whether the AI system is procured from a vendor, built internally, or sourced and adapted from a third party. This is a broader obligation than many entities have recognised. AI capabilities embedded in commercial software, AI tools adopted by individual business units outside formal ICT procurement, and AI features activated within existing platforms all fall within the scope of the requirement. An entity that has not inventoried its AI use cannot complete its self-assessment obligations.

The self-assessment addresses risk across several dimensions: privacy, security, transparency, explainability, and contestability. Each dimension requires the entity to identify the risks present in the AI system and the controls in place to manage them. The outcome of the self-assessment is a risk level for the system. Where the residual risk reaches a mid-range threshold or above, the project must be submitted to the AI Advisory Board for independent review. The same escalation requirement applies to projects financed through the Digital Capability Fund or with a total cost exceeding five million dollars. For entities with AI systems in this category, Advisory Board submission is a requirement, not an option.

The AI Register is the accountability mechanism at the portfolio level. The Office of Digital Government collates use cases from completed self-assessments into a central register of AI use across the WA public sector. This makes the governance obligation externally visible: entities cannot treat AI adoption as an internal matter if they are required to submit use cases to a central register. The register also creates a record that will be available to auditors, ministers, and the public over time.

The practical implication for WA public sector entities is that AI governance can no longer be deferred. Entities that have adopted AI tools without formal governance, that have not inventoried their AI use, or that have not nominated an AI Accountable Officer are already behind the requirement. The toolkit does not phase in these obligations: it establishes them as current requirements for all entities in scope. The starting point for most entities is an AI inventory: understanding what AI systems are in use, where they came from, what decisions they influence, and what risk profile they carry. Without that baseline, neither the self-assessment nor the Accountable Officer role can be fulfilled.