Australia's AI Review Committee sets out expert advice on public sector AI use

The Australian Government's Artificial Intelligence (AI) Review Committee has held its first meeting, providing expert advice to two agencies on upcoming AI initiatives and agreeing to the terms of reference that will govern its work.

Established under the AI Plan for the Australian Public Service (APS), the committee offers non-binding advice on selected AI use cases before deployment - including those assessed as high-risk, highly sensitive, novel or complex - and will review agencies' responses to serious AI incidents.

At its June meeting, the committee provided advice to the Attorney-General's Department and Services Australia.

The Attorney-General's Department discussed using AI to support staff productivity and wellbeing through document analysis, matter triage and workflow optimisation. Services Australia presented a proof of concept using AI to help staff locate internal operational documentation.

Committee members welcomed the focus on staff productivity while highlighting the importance of human oversight, quality assurance and monitoring. They also encouraged both agencies to maintain focus on privacy, workforce capability and planning for future scale.

The committee will review AI projects before deployment - assessing whether proposed solutions are appropriate, risks have been identified and safeguards are adequate - and will also consider alignment with Australia's AI Ethics Principles and broader government AI policies.

The terms allow the committee to draw on external experts where needed, provided conflicts of interest are managed and security requirements are met. External advisers may contribute perspectives but will not participate in developing the committee's final advice.

The committee's effectiveness will be reviewed at least annually.

"Members bring cross-disciplinary expertise to provide a strong foundation for rigorous, practical advice," said Chris Fechner, chief executive officer of the Digital Transformation Agency (DTA).

"This ensures the Committee's advice is grounded in how AI is actually deployed in government, and carries the authority needed to inform senior decision-making."

Fechner said public trust would be critical to the successful adoption of AI across government.

"That's why we are committed to being transparent about how the Committee operates and to continually reviewing and refining its approach as we learn from experience," he said.

The committee spans expertise across digital transformation, AI safety and assurance, privacy and data ethics, cybersecurity, and workforce and integrity policy.

Agencies remain accountable for implementation, decision-making and risk management.

The committee will meet as required to consider AI use cases and serious incidents referred by agencies.