Since 20 October 2026, anyone who uses generative AI to prepare a document for a Fair Work Commission case has to say so. That includes the HR adviser who drafted the employer's response between meetings. On paper the Fair Work Commission AI rules are a litigation matter. In practice they are an AI governance test for every HR team in the country.
This is general information and not legal advice. The detail below draws on HWL Ebsworth's 16 October 2026 summary of the Commission's guidance note.
What the rules actually require
The guidance note was finalised with a statement from Justice Hatcher on 24 August 2026. It applies to every participant in a case: applicants, employers, HR advisers, union officials and lawyers. In short:
- Anyone who uses generative AI to prepare a document lodged with the Commission must disclose that use, check the document and state that they have checked it.
- Using AI only for spelling, grammar or formatting is not covered.
- Witness statements must reflect the witness's own knowledge and words. AI can transcribe or translate speech but must not create the factual content.
- Parties should not put personal or confidential information into public AI tools.
- Professional representatives, including lawyers, paid agents and in-house HR advisers, must also hyperlink the case law they cite.
The Commission is also adding an AI disclosure section and a checking declaration to its forms in individual rights matters. Everyone completing those forms must complete both, whether or not they used AI.
Non-compliance can lead to a document being given less weight or disregarded, a costs order or dismissal of the case. HWL Ebsworth also notes that knowingly giving false or misleading information to the Commission is an offence.
Do I have to disclose AI use to the Fair Work Commission?
If you used generative AI to prepare a document you are lodging, yes. You disclose it, you check it and you say you have checked it. If AI only fixed your spelling or formatting, no. If the document is a witness statement, the content has to be the witness's own, whatever tool helped with the typing.
The part many miss is that the declaration applies to everyone completing the form, including people who did not use AI. That makes it a question every HR team will now be asked on every filing.
The Aldi case shows what unchecked AI costs
On 19 August 2026 the Commission ordered Sadnan Khan to pay $1,230 of Aldi's costs after he used AI to start and continue an unfair dismissal claim that could not succeed. He had not served the six months of continuous employment required. According to the CCIWA summary, Deputy President Easton warned him twice that the claim lacked merit, and his later AI-generated submissions cited the wrong termination date.
The Deputy President accepted that AI can be a useful drafting tool when used properly, but said the applicant was responsible for checking his eligibility and the accuracy of the material. HWL Ebsworth notes the Commission described such a costs order as very rare. What this shows is that the Commission is not punishing AI use. It is punishing unchecked AI use.
Where employer AI policies usually fall short
Most employer AI policies cover customer-facing chatbots and internal productivity tools. Very few mention external filings.
Consider a composite situation. An HR manager pastes performance review notes and a termination timeline into a public chatbot to draft a response to an unfair dismissal claim. The tool tidies the timeline and quietly shifts a date. The response is filed. Now there is an undisclosed AI use, confidential employee information in a public tool, and a factual error in front of the Commission. Three problems from one lunchtime shortcut.
An AI use policy for employers in Australia now needs to answer a few uncomfortable questions. Which tools can HR use for tribunal documents? Who checks the output against the source records? Where is the record that shows which tool drafted what?
What This Means for Your Organisation
What we see across organisations is that HR is often the least governed AI user in the business. The tools are free, the work is sensitive and the pressure is real. The Commission's rules do not create that risk, but they make it visible.
The practical fix is not a ban. It is a short, specific rule for AI in HR matters: approved tools only, no personal information in public tools, a named person who verifies facts against records and a simple log of AI use that feeds the Commission's disclosure.
Key Takeaways
- The Fair Work Commission AI rules apply from 20 October 2026 and require disclosure, checking and a checking declaration for AI-prepared documents.
- Spelling, grammar and formatting help is excluded, and witness statements must stay in the witness's own words.
- The Khan v Aldi costs order shows the Commission penalising unchecked AI, not AI use itself.
- Update your AI use policy to cover tribunal filings, approved tools, source verification and a record of AI use.
How Trusenta Can Help
- AI Governance Foundations builds the policies and decision rights that cover uses like HR filings, which most first-generation AI policies miss.
- AI Governance Register lets you record which AI tools are approved for which tasks and who is accountable for each.
- Compliance Management helps you track new obligations like the Commission's guidance note against the controls that satisfy them.
Conclusion
The Commission has not asked employers to stop using AI. It has asked them to stand behind what they file. Organisations that already know which tools their people use, and who checks the output, will find the new rules easy. The rest will find out how much they rely on AI the first time a form asks.
