AI in Fair Work matters is becoming more common, and it can create serious risks if used improperly.
Earlier this year we supported an employer with a matter involving an underperforming employee in an entry-level role. The employer had raised concerns about the employee’s work performance and sought a response. Nothing unusual there.
What was unusual was the response that came back.
Despite the employee’s written communication skills generally being quite limited, they produced a 20+ page document that looked, at first glance, like something prepared by a lawyer. It included multiple references to Fair Work cases, “judgements”, and complex legislation. As we took a deeper look at the response however, it became apparent that a number of the references included were plainly irrelevant, incorrect, or simply didn’t make sense in the context of the matters being addressed.
Even so, we still had to work through it carefully and help the employer respond. As the matter proceeded, it also became clear that the employee didn’t fully understand the content of their own response, which created a further layer of complexity. The matter was ultimately resolved, but the use of AI didn’t add much value for either party. Indeed, it added quite a bit of time, confusion, and cost.
That experience is not unique. Generative AI tools such as ChatGPT, Claude, Copilot and Gemini are now widely accessible, and people are using them to draft everything from general workplace correspondence through to grievances and complaints, unfair dismissal, general protections and other types of formal claims (and associated responses).
There’s no doubt that AI-technology is here to stay and, used well, we agree that it can be genuinely helpful. But used poorly, it can create serious risks, especially in legal or quasi-legal processes like Fair Work Commission matters.
Indeed, in a recent Unfair Dismissal case before the Fair Work Commission (FWC), the Applicant found out the hard way that AI may not always be the helper we might hope it to be.
In Sadnan Khan v Aldi Pty Ltd [2026] FWC 3144, the Applicant was ordered to pay part of ALDI’s (his former employer) costs when he, in the eyes of the Commissioner, acted unreasonably by persisting with a claim that had no chance of success despite him being repeatedly warned that was the case (he was ineligible to bring a claim because of his short period of employment). An important factor impacting the Commissioner’s order for him to pay some of ALDI’s costs was that his application was largely AI-generated and contained multiple errors which resulted in ALDI incurring additional costs which could have been prevented had the Applicant acted more reasonably.
The Fair Work Commission’s Guidance Note: a welcome (and necessary) starting point
In response to matters such as those above and the huge increase in the use of AI in Fair Work matters (a Fair Work Commission representative in June this year stated to a Senate committee that “40 to 50% of people were telling us that AI had been used to make an application to the Commission”), we were pleased to note that the Commission recently published a Guidance Note relating to the use of generative AI in Commission cases, together with practical resources explaining the benefits, risks and requirements relating to GenAI use in case matters.
In simple terms, the Guidance Note requires that, from 20th October 2026, if any party uses AI in Fair Work matters to help them prepare documentation, the fact that it has been used needs to be disclosed. You will also need to include detail of how it’s been used (for example, if it was used to; complete research and give advice, organise facts and evidence, transcribe audio statements, and/or write or edit content).
The Guidance Note also clearly reinforces that parties to an FWC matter remain ultimately responsible for the accuracy of all documents filed. That means you need to check facts, evidence, and citations, ensure the relevance and accuracy of your documents, and declare that you have completed those checks. You cannot make the assumption that because material to support your case has been AI-generated, that it is relevant and correct (regardless of how impressive it might appear).
The requirements outlined in the guidance note extend to both Applicants (usually employees or former employees) and Respondents (usually employers). Legal practitioners and other professional representatives of parties to FWC matters have an added responsibility to include hyperlinks to any case law referred to in documents prepared using GenAI.
The guidance note is a sensible step in our view and sets a clear tone: while AI can assist, it doesn’t replace careful human judgement or accountability.
Click HERE to access the Fair Work Commission’s Guidance Note and additional resources relating to the use of GenAI in FWC matters.
Of course, reach out if you need further information or assistance.
Note that this is general information only. It is not formal or legal advice. Information is current and we believe accurate at the time of publication – 16/09/2026.
