WA Health Artificial Intelligence Policy MP 0193/25 and Standard 8.8: Principle 8, accountability and records: oversee every AI system through a clear governance structure with defined roles, responsibilities, accountability and liability, identifying the AI Accountable Officer, an Executive Sponsor who is a Clinical Executive where patients are affected, the Accountable Supplier, Information Custodian, Authorised Officers, Technical Oversight, an AI Model Validator, an Independent AI Auditor where used, the clinician as final decision-maker, developers and data staff who assess the system under documented plans, and users who stay within the approved use case, train, monitor output quality and report incidents; and keep all records created by or about AI, including design, testing, implementation, evaluation and decommissioning, in the approved recordkeeping system under the State Records Act 2000 following the State Records Office AI recordkeeping advice

What does 8.8 require?

WA health entities must ensure AI systems are overseen by a clear governance structure with defined roles and responsibilities, accountability and, where relevant, liability. The primary roles to identify are the AI Accountable Officer, an executive overseeing AI strategy and governance; the Executive Sponsor with overall accountability for the system meeting requirements, who must be a Clinical Executive if the system affects patients; the Accountable Supplier, the executive overseeing development whether in-house or by a vendor; the Information Custodian responsible for data governance of the information asset; Authorised Officers acting under the entity's authorisations and delegations schedule; Technical Oversight by an executive technical officer such as the chief information officer; an AI Model Validator who evaluates performance for explainability and robustness; an Independent AI Auditor who assesses conformance with policy, standards and law; the clinician as final decision-maker where AI informs clinical decisions; designers, developers, data engineers, analysts and scientists who continuously, systematically and transparently assess the system under documented monitoring, testing and evaluation plans; and users and end users who use the system only within the approved use case, mandatory policies and law, complete required training, continuously monitor the quality and accuracy of outputs and raise incidents or risks through the organisation's reporting processes. Entities must also keep all records created by or through AI in accordance with the State Records Act 2000, retain records of AI use in business activities including design, testing, implementation, evaluation and decommissioning in the approved recordkeeping system, and follow the State Records Office advice on AI and recordkeeping.