AI Regulation Tracker / Tribunal guidance

Australia's Fair Work Commission drafts rules requiring parties to disclose and verify AI-prepared filings

The Commission published an exposure-draft Guidance Note that would require anyone using generative AI to prepare a filing to say so and to confirm they checked every fact, case, and quote. It is a draft open for comment, not a final rule.

By Anthony Guerriero · Reviewed by The Leveraged Years Editorial Desk · Published July 7, 2026 · Last updated July 7, 2026

Australia's Fair Work Commission drafts rules requiring parties to disclose and verify AI-prepared filings regulation briefing

The Leveraged Years AI Regulation Tracker

Regulatory Briefing

Bindingness: Regulatory Guidance
Scope: Australia, national workplace relations system. Parties, lawyers, and paid agents filing in Fair Work Commission cases, including unfair dismissal and general protections matters.
What Changed: The Fair Work Commission released a President's statement and an exposure-draft Guidance Note, "Use of Generative Artificial Intelligence in Commission cases," dated March 24, 2026. The draft sets three requirements: disclose that generative AI was used, verify all details against reliable sources and state that this was done, and ensure any witness statement or declaration reflects the witness's own knowledge.
Who This Affects: Any person who uses generative AI to write, create, modify, or otherwise prepare an application or other document lodged in a Commission case, including self-represented litigants, employment lawyers, and paid agents.
Current Status: Proposed. Exposure draft published for consultation; commencement date left as "[insert] 2026" in the draft.
Effective Date: Not yet commenced. The comment period closed 4 pm, Friday, April 10, 2026; the guidance remains an exposure draft.
Primary Source: Fair Work Commission, "President's statement and draft guidance for use of generative AI published," https://www.fwc.gov.au/about-us/news-and-media/news/presidents-statement-and-draft-guidance-use-generative-ai-published
Secondary Analysis: Fair Work Commission, "Guidance Note: Use of Generative Artificial Intelligence in Commission cases" (exposure draft, published March 24, 2026).
Workflow Impact: Practitioners must add an AI-use statement to affected filings and run an independent citation and fact check before lodging, because the draft bars using AI itself to confirm accuracy.
Operational Next Step: Read the exposure draft and map its three requirements against your current filing checklist, because the exposure draft signals the verification standard the Commission expects.
Last Verified: July 7, 2026

Australia's Fair Work Commission has published an exposure-draft Guidance Note that would set clear obligations on anyone who uses generative artificial intelligence to prepare a filing in a Commission case. The document, "Use of Generative Artificial Intelligence in Commission cases," is dated March 24, 2026 and was released with a statement from the Commission's President. It is a draft open for consultation, not a final rule, and the Commission has left its commencement date blank pending the outcome of comments, which closed at 4 pm on April 10, 2026.

Three requirements for AI-assisted filings

The draft sets three obligations. First, if generative AI is used in writing, creating, modifying, or otherwise preparing any document to be lodged, the filer must state in that document that AI was used. The draft offers model wording: "GenAI was used in preparing this document." Second, the filer must check the document and correct it so that all details are accurate and relevant, and must confirm in the document that this was done. That check must ensure, in the draft's words, that "all cases, legislation, textbooks and articles referred to in the document exist and stand for the legal positions attributed to them," that facts and evidence cited are correct and real, and that quotes are accurate and properly attributed. Third, any witness statement or declaration must reflect the witness's own knowledge, and the witness must declare that it does. The draft goes further on this point, recommending that filers not use generative AI to create the content of witness statements or declarations at all, since those documents must reflect a person's own account rather than machine-generated text.

Verification cannot be outsourced to the AI

A central feature of the draft is that a filer cannot verify accuracy by asking a chatbot. The Commission states that checking must be done "by looking at sources of information known to be correct," and it names them: Commission benchbooks and legal texts for principles, the Commission's Find decisions and orders service, AustLII for court decisions, and the Federal Register of Legislation for Commonwealth statutes. The draft is direct about why. Generative AI, it says, "may make up fake cases, citations and quotes, or refer to laws, articles or legal texts that do not exist," and may cite decisions from other jurisdictions or ones that have since been overturned.

The context, and one figure to treat with care

The Commission has framed the guidance as a response to a sharp, AI-driven rise in its workload, including filings that contain fabricated or hallucinated legal claims. Reporting on the rollout has cited a caseload surge of roughly 70 percent attributed to generative AI use. That percentage should be treated as reported rather than as a confirmed Commission statistic, because the figure did not appear in the exposure draft or the primary announcement pages reviewed here. The broader point is not in dispute: the Commission says AI-generated material "may be inaccurate, incomplete, out of date or just made up."

What non-compliance costs

The draft carries real consequences. It warns that failing to meet the requirements "may affect your case," including documents being given less weight or disregarded, an order to pay another party's costs, or dismissal of the case. It also flags a separate criminal exposure. Knowingly giving false or misleading information to the Commission is an offence under section 137.1 of the Criminal Code, punishable by up to 12 months' imprisonment. The draft further cautions filers not to feed personal, confidential, or legally privileged information into public AI tools, noting that such data may be retained and could become public.

What the draft does not do

The Guidance Note does not ban generative AI. It acknowledges that AI can improve efficiency and can help self-represented litigants express themselves more clearly. It does not impose new penalties beyond those already available to the Commission, and it does not yet bind anyone, since it has not commenced. For a US reader, the Commission's rules do not reach filings made to US agencies or courts. They would apply to any employer, in-house team, or firm lodging documents in an Australian Fair Work matter, and they track a wider pattern of courts and tribunals worldwide requiring disclosure and human verification of AI-assisted submissions. That direction is worth watching for anyone advising across borders. The practical lesson travels well beyond Australia: where a filer uses AI, the burden of confirming that its citations and facts are real sits with the human, not the tool.

Frequently Asked Questions

What did the Fair Work Commission actually change?

Nothing is binding yet. The Commission published an exposure-draft Guidance Note dated March 24, 2026 that would require filers who use generative AI to disclose that use, independently verify every fact and citation, and confirm in the document that they did so. Its comment period closed April 10, 2026, and it remains an exposure draft.

Who does the draft apply to?

Anyone who uses generative AI to write, create, modify, or otherwise prepare an application or other document lodged in a Commission case. That includes parties, self-represented litigants, employment lawyers, and paid agents in matters such as unfair dismissal and general protections.

Can I verify AI-generated citations by asking the AI to check them?

No. The draft is explicit that checking must be done against sources known to be correct, such as AustLII, the Federal Register of Legislation, and the Commission's own benchbooks and decisions database, not by asking the AI tool.

What happens if I ignore the requirements?

The draft warns that your documents may be given less weight or disregarded, that you may be ordered to pay another party's costs, or that your case may be dismissed. Knowingly giving false or misleading information can also be an offence under section 137.1 of the Criminal Code, carrying up to 12 months' imprisonment.

Does this ban the use of generative AI in Commission cases?

No. The draft permits AI use and notes its efficiency and access-to-justice benefits. It requires disclosure and human verification rather than prohibition, and it recommends against using AI to generate the content of witness statements or declarations.