24 September 2026
11 min read
#Workplace Relations & Safety, #Digital Economy, #Government
Published by:
A massive 70% increase in the total overall workload of the Fair Work Commission (FWC) over the past three years, largely attributed to the proliferation and use of AI in unfair dismissal and general protections applications, has led to the publication of a new guidance on the use of generative artificial intelligence (GenAI) in Commission cases (AI Guidance Note).
Effective from 20 October 2026, the AI Guidance Note introduces new disclosure and verification requirements for parties using AI-generated materials in proceedings and makes clear that responsibility for the accuracy, relevance and legal soundness of any application or document remains with the person preparing or lodging the material.
The guidance follows the recent FWC decision in Sadnan Khan v Aldi Pty Ltd [2026] FWC 3144 (Aldi decision), in which the FWC reinforced that parties may face adverse costs orders if they fail to use AI reasonably, including by failing to identify important jurisdictional errors in AI-generated material.
The AI Guidance Note sets out three key requirements that apply if a party uses GenAI to help prepare any document to be lodged in a FWC case, including an application or employer response. In summary, parties must:
The AI Guidance Note also cautions against entering any personal, confidential or legally privileged information into GenAI, and highlights the potential risk for parties breaching FWC orders prohibiting or restricting the publication of certain information if that information is entered into the tool.
In addition, the FWC plans to introduce modified forms giving effect to the guidance and requiring, among other things, parties to disclose where AI has been used in the preparation of documents lodged with the FWC. Both the AI Guidance Note and the modified forms will take effect on the same day.
Failure to comply with the AI Guidance Note may result in documents being given less weight or disregarded, an adverse costs order being made against the relevant party or the party’s case being dismissed altogether.
The AI Guidance Note also reiterates that GenAI is not a substitute for legal advice and should not be relied on to assess the merits of a claim or a party’s prospects of success.
The AI Guidance Note has been released in the broader context of increased use of, and reliance by parties on, GenAI to generate materials for use in FWC processes. As President Hatcher noted in his statement delivered in conjunction with the AI Guidance Note, the proliferation of matters commenced in the FWC has been attributed to the growing use of AI.
The FWC has also observed changes in the character of materials lodged by parties, including longer and more detailed written submissions, greater use of formal or legalistic language, and more structured claims by self-represented applicants.
The GenAI Guidance note is intended to address some of the identified problems reliance on GenAI can create for parties to a case and the FWC, including unrealistic expectations about prospects of a claim and likely compensation, as well as inaccurate, incomplete, out of date and made up AI generated content, which can impose a significant time and cost burden to parties and the FWC.
The FWC are not alone in responding to these challenges. Courts and tribunals at the federal, state and territory levels are grappling with the increased reliance on GenAI and the associated risks to the proper administration of justice and public confidence in the legal system.
The recent FWC decision of Sadnan Khan v Aldi Pty Ltd [2026] FWC 3144 demonstrates the possible implications for parties that fail to use AI in a reasonable manner, including adverse costs orders being made against them.
Background
In this decision, the FWC ordered the applicant to pay part of the respondent’s costs due to his inappropriate use of AI.
Mr Khan commenced employment with ALDI on 29 September 2025 and was notified of his dismissal on 26 March 2026, three days short of the six-month minimum employment period for protection from unfair dismissal required under sections 382 and 383 of the Fair Work Act 2009 (Cth) (FW Act). Later that evening, Mr Khan lodged an unfair dismissal application which, on its face, was prepared using GenAI and contained visible AI drafting notes.
After receiving the application, Commission staff emailed Mr Khan to advise that he appeared not to have served the minimum employment period for protection from unfair dismissal. They explained that the relevant date under section 383(a)(i) of the FW Act was the date he was notified of his dismissal, not the date the dismissal took effect. Mr Khan’s application was then referred to Deputy President Easton’s Chambers to determine whether the minimum employment period was completed.
On 31 July 2026, Deputy President Easton’s Chambers wrote to Mr Khan explaining why his case was unlikely to succeed due to the operation of section 383 of the FW Act, and warned him of the risk of an adverse costs order against him if he continued with the application. Mr Khan then sent an AI-generated reply to the Deputy President, which again addressed the wrong end date – the date the dismissal took effect rather than the date he was notified of the dismissal. His response also failed to address the matters raised by the Deputy President. ALDI then indicated it would seek costs if the claim proceeded to hearing.
Later that day, Deputy President Easton’s Chambers emailed Mr Khan again to explain that there was a strong likelihood his case would not succeed and that there was a very real possibility the Deputy President could make an order requiring him to pay ALDI’s costs. Despite this further warning, Mr Khan continued to rely on the same AI-generated arguments.
Only at the hearing, after being taken through section 383 of the FW Act, did Mr Khan admit that he had not served the minimum employment period and discontinued his application. Considering Mr Khan’s conduct, Deputy President Easton determined it was appropriate to order him to contribute to ALDI’s legal costs for preparing and attending the hearing.
Basis for the costs order
Under section 400A of the FW Act, the Commission may order a party to pay another party’s costs where it is satisfied that an unreasonable act or omission of that party, in connection with the conduct or continuation of the matter, caused the other party to incur those costs. This requires two elements to be satisfied:
Deputy President Easton found both elements were satisfied.
Mr Khan’s decision to continue his claim after 31 July 2026, when he had been advised by the Deputy President that there were no reasonable prospects of success, was unreasonable. The Deputy President emphasised that although Mr Khan had no legal training, the FWC’s processes are designed to cater for self-represented litigants and that Mr Khan did not need legal training to understand the Commission’s emails.
The Deputy President was also satisfied that, if Mr Khan had properly read the emails from the FWC and the words of section 383 of the FW Act, he would have known that he had not served the minimum employment period and was therefore not eligible to make an unfair dismissal claim. Mr Khan’s decision to continue his claim despite receiving clear information from the Commission was, for the purposes of the costs provision of the FW Act, unreasonable.
In respect of the second element, ALDI only sought to claim for costs incurred from 31 July 2026, being the costs of preparing for and attending a hearing that need not have occurred. Exercising the resulting discretion, the Deputy President ordered Mr Khan to pay $1,230, calculated in accordance with the maximum rates recoverable under Schedule 3.1 of the Fair Work Regulations 2009 (Cth).
What does the decision say about AI use in FWC proceedings?
This decision treated Mr Khan’s GenAI use as a factor in the reasonableness assessment. However, it did not specifically penalise Mr Khan for using GenAI.
Deputy President Easton stated that “[u]sing AI to assist in drafting correspondence and even submissions is not, of itself, unreasonable. Artificial intelligence, in 2026 at least, is a drafting tool that can assist litigants when used properly”. What proved decisive was Mr Khan’s continued, unquestioning reliance on the AI-generated output after he had been repeatedly warned by the FWC that his arguments were incorrect.
The Deputy President observed that it was “quite possible that Mr Khan adopted AI as his quasi-legal advisor and then blindly submitted AI generated outputs as his own material” without checking them against the correspondence he had received or publicly available FWC information. It was Mr Khan’s responsibility to determine whether he was eligible to make the claim and to ensure that anything he submitted to the FWC addressed the concerns raised.
Furthermore, the decision distinguished between using AI as a drafting aid, which the Deputy President did not criticise, and relying on AI-generated content without exercising independent judgment, particularly in the face of repeated warnings that an individual without legal training could readily understand.
The FWC’s emphasis on the accessibility of its own communications and published guidance suggests that tribunals may expect self-represented litigants to test any AI-generated content against those materials. The Deputy President stated that “[i]f Mr Khan had properly read for himself the AI-generated material on which he relied so heavily, and compared it to the general information on the Fair Work Commission’s website, and compared it to the specific information he had been provided from the Commission member deciding his case, he would have realised that his AI arguments were just plain wrong”.
In deciding costs, the Deputy President also noted the need to deter not only Mr Khan, but also other applicants in the same position, from using GenAI in an unreasonable manner.
The AI Guidance Note applies to all parties, not just applicants. From 20 October 2026, parties lodging documents in FWC proceedings will need to disclose when and how GenAI has been used.
The note also makes clear that responsibility for the accuracy, relevance and legal soundness of a document remains with the person who prepares, lodges or adopts it. AI-generated facts, authorities, quotations and submissions cannot be accepted without appropriate verification.
While the FWC does not criticise the use of GenAI, the Aldi decision demonstrates the risks of relying on AI-generated content without exercising independent judgment. Parties may face adverse costs orders, have documents given less weight or disregarded, or, in some cases, have their claims dismissed. The AI Guidance Note also notes the potential criminal implications of providing false or misleading information to the FWC, which may breach section 137.1 of the Criminal Code Act 1995 (Cth).
The Aldi decision is significant in circumstances where both limbs of the test under section 400A of the FW Act are required to be satisfied for the FWC to exercise its discretion. Historically, there has been a high threshold to satisfy both limbs of the test.
Parties planning to use GenAI should also be aware that the guidance and requirements attached to the use of the technology in proceedings vary between different courts and tribunals. For instance, the NSW Industrial Relations Commission adopts a much stricter approach by prohibiting the use of GenAI in the preparation of statements and evidence, and requiring leave to be granted to use GenAI in preparing or generating annexures or exhibits to an affidavit or statement.
If you have questions about the new AI Guidance Note or need assistance reviewing your organisation’s AI policy, please get in touch with our team.
Disclaimer
The information in this article is of a general nature and is not intended to address the circumstances of any particular individual or entity. Although we endeavour to provide accurate and timely information, we do not guarantee that the information in this article is accurate at the date it is received or that it will continue to be accurate in the future.
Published by: