1. AI-assisted claims and the unpredictability of outcomes
The recent Fair Work Commission (FWC) decision in Baker v Macquarie University [2026] FWC 3054 involved a casual academic who succeeded in a claim against his university employer. Mr Baker was self-represented and used premium AI tools, including ChatGPT Pro and Claude, to prepare and conduct his case. The decision was the first in the country to apply the employee choice casual conversion provisions.
One week later, in Khan v Aldi [2026] FWC 3144, a dismissed Aldi worker who similarly relied on AI for an unfair dismissal claim was ordered by the Commission to pay costs. The Commission characterised his argument as 'just plain wrong' and made him personally liable for $1,230 of his employer's legal costs, a notable outcome given how uncommon costs orders are in the Commission.
2. The Fair Work Commission caseload
The Fair Work Commission has experienced a marked increase in lodgements in recent years. Annual lodgements rose from just over 30,000 in 2022-23 to approximately 40,000 in 2023-24, exceeded 44,000 in 2024-25, and are projected to reach between 50,000 and 55,000 in 2025-26, representing growth of more than 40% over a three-year period.
3. The consequences of AI misuse by applicants
On the applicant side, recent decisions illustrate the range of adverse consequences of AI misuse:
- In Reece Hoverd v M & J D Pty. Ltd. [2026] FWC 1013, the AI-assisted applicant relied on contract and award clauses that did not exist. The claim was dismissed and the employer was invited to apply for costs.
- In Sarah Byrne v CXC Corporate Services Pty Ltd [2026] FWC 2751, the applicant declared in writing that she had verified her AI-assisted submissions. The Commission found this declaration to be misleading. The applicant had not read one of the authorities she cited, and one of the cited cases did not exist.
- In Application by Pennisi [2026] FWC 352, the difficulty was not fabrication but volume. The application contained extensive repetitive AI-generated submissions with shifting reasoning and internal inconsistencies. The Commission rejected the application for an extension of time.
4. The consequences of AI misuse by employers
Employers and their advisers are equally susceptible to adverse outcomes arising from uncritical reliance on AI.
- In Hayley Lord v Millet Hospitality Geelong Pty Ltd [2025] FWC 2740, the employer used ChatGPT to draft a redundancy notification, which was sent to the affected employee by email after business hours. The notification failed to discharge the applicable award consultation obligations. The Commission was highly critical of the approach, describing it as falling short of 'basic standards of decency'. As a result, the Commission found the redundancy was not genuine and awarded the employee one week's compensation.
- In Daniel O'Hurley v Cornerstone Legal WA Pty Ltd [2024] FWC 1776, a law firm in Western Australia used ChatGPT to draft correspondence intended to record an employee's abandonment of employment. The Commission interpreted the correspondence as a notice of termination, which defeated the employer's own jurisdictional objection. The employer could not argue that the employee had abandoned their role, and the general protections claim was allowed to proceed.
5. Procedural reforms under the Building Cooperative Workplaces Act
The Workplace Relations Legislation Amendment (Building Cooperative Workplaces No. 1) Act (Cth) commenced on 7 July 2026, introducing changes to the Fair Work Commission's procedural framework. While the Act does not alter the underlying legal tests governing unfair dismissal and general protections, it is intended to streamline the management and progression of matters before the Commission. Four reforms are particularly noteworthy.
- Removal of the automatic jurisdictional gateway. Where there is a dispute as to whether a dismissal has occurred, the Commission was previously required to determine that jurisdictional issue before taking further procedural steps. The Act removes that requirement. While jurisdictional objections remain available, the Commission may now proceed directly to conciliation, increasing the likelihood that parties will engage in settlement discussions at an earlier stage of the proceeding.
- Determination on the papers. The Commission now has broader discretion to determine matters on the basis of written submissions alone, with the consent of the parties. This option extends to unfair dismissal matters and may offer a streamlined pathway for employers with strong documentary records. However, it may present challenges where the written record is incomplete, as parties will be unable to address evidentiary deficiencies through oral testimony.
- Streamlined general protections certificates. A broader range of Commission personnel may now issue the certificates required for general protections matters to proceed to the Federal Circuit and Family Court or the Federal Court. This reform is intended to accelerate the progression of unresolved general protections disputes beyond the Commission.
- Vexatious litigant orders. The Commission may now make orders preventing repeat applicants from bringing further proceedings where it has previously struck out an application as frivolous, vexatious or without merit. It is advisable to verify whether a prospective applicant has been subject to such an order, as it may affect the management and costs of the matter.
Taken together, these reforms reflect a broader policy objective of expediting the resolution of workplace disputes.
6. New FWC AI guidance taking effect on 20 October 2026
The Commission published its final AI Guidance Note on 24 August 2026, with the new obligations commencing on 20 October 2026. The guidance applies to any document lodged in a Commission proceeding, irrespective of the type of application.
There are three principal requirements:
- Disclosure. Parties must state in the document that GenAI was used and how they used it.
- Verification. Every fact, authority, quotation and statutory provision must be verified for accuracy and existence. The lodged document must also state that verification has occurred.
- Authenticity. Witness statements and declarations must genuinely reflect the witness's own knowledge, own words, and must be true to the best of their knowledge.
The Commission has framed this guidance as an initiative to encourage responsible AI use rather than as a punitive measure. Professional representatives, including in-house HR professionals appearing for a party, must also hyperlink all case law cited. The Commission is likely to hold professionals to a stricter standard than self-represented parties.
7. Practical steps to implement before October
With the Fair Work Commission's AI guidance taking effect in October, employers should consider taking practical steps to mitigate risk and ensure compliance.
- Establish a formal AI disclosure and verification protocol. Any person responsible for drafting material to be lodged with the Commission, including claim responses, investigation reports and witness statements, should follow a carefully documented process for AI disclosure and verification.
- Close the file at dismissal. Decision-maker statements, performance records, consultation records and contemporaneous reasons for dismissal have always been critical in defending claims. However, with the Commission now empowered to determine matters on the papers, the quality and completeness of an organisation's documentary record will assume even greater importance.
