Refusal and review
The Tribunal now has a practice direction on generative AI, and it applies to cases already on foot
The President of the Administrative Review Tribunal made a Generative AI Practice Direction on 20 August 2026. It takes effect from the day it was signed and applies to every application, whenever it was lodged.
The Practice Direction is made under section 36(1) of the Administrative Review Tribunal Act 2024 by the President, the Hon Justice Emilios Kyrou AO, and is dated 20 August 2026. It records that it was adapted from the Federal Court’s Use of Generative Artificial Intelligence Practice Note (GPN-AI) of 16 April 2026, with the Chief Justice’s permission. It applies to parties and to their representatives, legal or non-legal, so a registered migration agent appearing in the Tribunal is covered as much as a lawyer.
It commences and has effect from the date it is signed
and applies to all applications whether lodged before, on or after this date
.
What is expected of a party or representative
Paragraph 2.4 sets out the Tribunal’s expectations. A person using generative AI is to have a basic understanding of its capabilities, limitations and risks. Its use must not adversely affect the administration of justice
, and the Direction says there will be circumstances where using it is inappropriate.
Three of the expectations bear directly on how a matter is run:
- if the Tribunal requires it, a person
must disclose to the Tribunal if (and how) Generative AI has been used in a proceeding and provide certification if requested
; - a person
must be able to explain, if the Tribunal asks, how Generative AI was used in connection with a proceeding, including what was done to check the output
; - a person
must not rely only on a Generative AI tool to verify that tool’s own output
.
Responsibility does not move. A person who uses generative AI must verify Generative AI-assisted material and remains responsible for it, whether or not Generative AI was used to prepare it
.
Statutory declarations and evidence
Paragraph 3.9 goes to the heart of a review file. Any use of generative AI must be consistent with the requirement that when a person makes a statutory declaration or witness statement, they are representing that the document reflects their own recollection, knowledge, opinion and/or experience
. The Direction points to the criminal laws that prohibit falsifying or interfering with evidence.
Confidential and restricted material
Paragraph 3.17 warns that information given to a publicly available tool may become available to other people
, and that users may not know where it is stored, how it is used, or who will have access. Paragraph 3.18 lists categories the law restricts, including material subject to a Tribunal confidentiality, non-publication or non-disclosure order, material protected by privilege, and material subject to the implied obligation not to use it for a purpose other than the proceeding it came from. Documents produced under sections 23 and 25 of the Act, under a Tribunal direction, or in answer to a summons are given as examples.
Sources
- Administrative Review Tribunal (Use of Generative AI) Practice Direction 2026 (PDF)
- Administrative Review Tribunal, practice directions and other guidance
- Administrative Review Tribunal Act 2024, section 36
Links go to the instrument or the official page. Where a page states a figure, the figure above is that page’s figure on the date shown.
Full guide: Refusal and review