AI is quickly becoming the new normal for the Fair Work Commission (Commission), prompting Justice Hatcher, President of the Commission to publish a statement[1] and an exposure draft of the proposed Guidance Note: Use of Generative Artificial Intelligence in Commission cases (GenAI Guidance Note)[2]. The GenAI Guidance Note comes after growing concerns that the Commission is experiencing an unprecedented increase in its workload, driven by applicants using artificial intelligence (AI) tools to prepare claims.
The statistics are alarming. The Commission received 44,075 lodgements in total across the 2024/25 financial year. As Justice Adam Hatcher noted in his presentation to the Victorian Bar Association in late February, the “normal” up until 2023 was just above 30,000 lodgements a year.
For 2025-26, projections sit between 50,000 and 55,000 lodgements. General protections and unfair dismissal claims have been identified as the key areas of growth, with an increased volume of claims, many with little merit. Notably, the historical correlation between dismissal-related applications and the general state of the labour market is no longer clear. The growth is principally due to the increasing use of AI tools by potential litigants.
AI tools can provide potential litigants with a false sense of understanding of the merits of their case, leading to claims being lodged with little to no legal basis. In particular, AI tools can:
In addition to misrepresenting a case, these limitations pose a significant risk of causing litigants to pursue time-consuming and costly claims, with little to no prospect of success.
The Commission has already seen what this looks like. In Reece Hoverd v M & J D Pty. Ltd[3], the applicant used AI tools to organise and draft submissions for a general protections claim. Deputy President Nicholas Lake found the applicant “consistently relied upon provisions of his contract and the Waste Management Award 2020 which did not exist as the basis for his argument”.
Even after being warned by the Commission not to provide false or misleading AI-generated evidence, the applicant continued with these submissions. The application was dismissed – and significantly, the Commission invited the employer to seek costs. If pursued, this would be the first costs order of its kind, meaning applicants may be significantly worse off when relying on unchecked AI outputs.
The increase in the volume of cases, often without merit, is only part of the concern. The bigger concern is what this volume of poorly articulated and plainly incorrect materials is doing to the Commission itself.
AI-assisted drafting has made it easier for parties to generate expansive narratives, multiple articulated grounds and extensive annexures. In turn, this requires lawyers to spend significantly more time reviewing, verifying and responding to material that is often repetitive, unfocused or legally misconceived.
As a result, matters that might once have been resolved early, or dealt with summarily, now demand greater upfront strategic assessment and more detailed written work. This elongates the lifecycle of proceedings, potentially increasing costs and the time it can take for a fair resolution to be identified.
For the first time in many years, the Commission has been unable to reach a key performance benchmark requiring, in 90% of cases, reserved decisions to be issued within 12 weeks. Year to date, the Commission has reached only 85%[4].
Of greater concern to President Hatcher is the impact the increased number of lodgements has on the Commission’s ability to deal with major cases. For example, a number of upcoming cases require the Full Bench of three or five members engaging in lengthy matters that are in considerable public interest. This includes reviews of award provisions relating to working from home and part-time employment. Undoubtedly, increases in meritless claims takes time away from such matters.
The Commission also has a core function to facilitate enterprise bargaining, and the pressure from increased general protections claims and dismissals directly reduces the resources available to do this – potentially meaning both workers and employers can be worse off.
The GenAI Guidance Note outlines three requirements that apply when a person uses GenAI in preparing any document (including an application) for the Commission:
Legal practitioners face an additional obligation to include hyperlinks to all case law cited. The Commission cites that the reason for doing this is to ensure that applicants remain accountable for the content of the documents they file.
The scale of the Commission’s workload pressures has also been recognised at a policy level. In the recent 2026–27 Federal Budget, the Government confirmed allocation of $1.3m to the Commission to continue providing specialised support to small businesses navigating disputes and Commission processes. This reflects an acknowledgement that increased lodgements are placing sustained pressure on the Commission’s dispute resolution framework. For workers and employers, this underscores the likelihood of ongoing delays, tighter procedural oversight and a continued focus by the Commission on managing its caseload efficiently.
Further, on 16 April 2026, the Federal Court of Australia (Federal Court) also released a Practice Note on the use of generative AI[5]. The Practice Note states that it is “fundamental to the administration of justice that Generative AI must be used appropriately, responsibly and with due care.”
The Federal Court now requires parties to disclose whether (and how) they have used generative AI in a proceeding. Further, like the GenAI Guidance Note, the Federal Court’s practice note also highlights that the generative AI may produce results that are not accurate, fictitious or plainly wrong, while reminding users of the risk they may breach confidentiality or privacy obligations if they do not carefully consider how they use these tools.
The Commission and Federal Court both raise key risks, and propose sensible mitigation measures proposed by the Commission, but they address the quality of what is filed without addressing the volume or other impacts to litigants. For example, a litigant who uses a generative AI model to draft a claim, and who then checks the citations and confirms the facts, still has access to the Commission (or a court), even if the underlying grievance has no reasonable prospect of success. As a result, requiring higher ‘quality’ submissions may do little to assist in tackling the challenge of a high volume of AI-assisted requests.
Disclosure and verification requirements are a necessary first step. Whether they are a sufficient one is the question the Commission, and the government, will need to confront sooner than they might expect.
If you found this insight article useful and you would like to subscribe to Gadens’ updates, click here.
Authored by:
Dudley Kneller, Partner
George Haros, Partner
Chris Girardi, Associate
Bailiejean Hohnberg, Lawyer
Luke Brusco, Lawyer
[1] President’s statement: Exposure draft of the Commission’s Guidance Note: Use of Generative Artificial Intelligence in Commission cases—opportunity to comment
[2] Guidance Note: Use of Generative Artificial Intelligence in Commission cases.
[3] [2026] COMMISSION 1013 at paragraph [5].
[4] Ibid.
[5] https://www.fedcourt.gov.au/law-and-practice/practice-documents/practice-notes/gpn-ai