In two recent cases both the Fair Work Commission (FWC) and Federal Circuit and Family Court of Australia (FCFCoA) have fought back against the indiscriminate and misguided use of artificial intelligence (AI) by applicants.
Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245
In Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245, Judge Symons of the FCFCoA dismissed an application in which the Applicant had, so her Honour inferred, inappropriately used AI to produce an originating process document that was integral to the case being advanced.
Judge Symons sets the scene at paragraph one:
“What follows is a cautionary tale about the dangers of befriending Artificial Intelligence (AI)-powered chatbots who masquerade as legal advisors.”
At the heart of this decision was a repeated failure by the Applicant to produce documents in an Evidence List (comprising three schedules setting out a total of 38 documents) that was attached to his Form 2 Application.
In response to the court’s suspicion that the Applicant may have been using generative AI and had failed to disclose it in accordance with the relevant FCFCoA practice note, the Applicant responded, “…all of the decisions are made by myself”. Judge Symons, however, drew an inference that, “…the Evidence List was generated by an AI tool of some description, and the documents referred to do not exist and/or do not bear the description that they have been assigned, making them impossible to identify.”
The Respondent, in response to the repeated failure of the Applicant to produce the documents listed in the Evidence List upon which he was relying, sought that the Application be dismissed. Judge Symonds granted this relief, stating “I had little hesitation in acceding to this application.”
At paragraph 52 Judge Symons acknowledges that AI, used appropriately, can have a constructive role to play in litigation:
“The Court acknowledges the potential appeal of a generative AI tool, such as a chatbot, to a self-represented litigant and especially one who does not speak the language adopted in the Court as his first language. The Court also acknowledges that responsible use of generative AI tools can have significant benefits for efficiency and access for justice; indeed, this is a cornerstone of the Court’s Practice Direction: Use of Artificial Intelligence that was issued on 29 May 2026. Without being exhaustive or prescriptive about its anodyne and helpful uses, AI tools can be used to improve the efficiency of routine or administrative tasks by, for example, organising material that already exists or summarising information that has been generated independently of the AI system. It might also usefully and as a starting point, identify legal information, such as whether time limits apply for the making of a particular claim. Where this information is capable of independent verification, it can assist in the demystification and accessibility of legal avenues for redress.”
Judge Symons did, however, strike a more sobering note about AI at paragraph 53:
“However, there are significant difficulties when AI answers a job description for legal advisor. This is because AI is not a lawyer, does not hold a practising certificate and is not authorised in any jurisdiction in Australia to engage in legal practice and to provide legal advice.”
Her Honour continued her critique of AI being used as a de facto lawyer at paragraph 55:
“…because AI is not a legal advisor, it owes no duty to the Court or the administration of justice and owes no duty to act in the best interests of Mr Ba or to observe regulatory or ethical standards, including in its engagement with other practitioners and parties. In other words, any AI tool used by Mr Ba did not care as to the truth or otherwise of any material that it might have generated. The sycophantic tendency of large language models to prioritise user approval over truth, can make them especially dangerous and unpredictable unless careful oversight, calibration and verification is applied.”
Judge Symons ended the decision, in which the application was dismissed, by noting at paragraph 59:
“There is one final observation that I wish to make about this case. Although both respondents deny, for different reasons, that Mr Ba’s case has any merit, the consequence of his decision to place blind faith in his AI “legal expert” means that he will not have the benefit of the Court’s adjudication of this question. While AI has the capacity to flatter and beguile, it also has the capacity to obscure and undermine. It would be an unsatisfactory outcome indeed if the misuse of AI resulted in a good (or at least arguable) claim being thrown on the proverbial scrap heap.”
Ms Wayan Buschman v Good to Great Schools Australia (C2026/5556)
In the case, Ms Wayan Buschman v Good to Great Schools Australia (C2026/5556), Deputy President Dobson of the FWC, in making orders dismissing a general protections application dealing with a dismissal, noted the counterproductive role of AI in the case observing:
“In short I found that the Applicant was engaged by Upwork, a Digital Labour Platform in the United States of America and therefore the matter was outside the jurisdiction of the Fair Work Act and the Fair Work Commission. I note that this was a case where artificial intelligence (AI) was used to prepare the Applicant’s case. Case law was incorrectly applied and gave false hope as to the merits of the case. The Applicant certified that they had checked the AI work carefully however during the proceedings it became clear that the Applicant had not thoroughly read or understood the case law instead only reading ‘the excepts presented by the AI”. This resulted in the Applicant bringing a case that should not have been brought creating unnecessary cost to the Tribunal and the Respondent.”
Some Observations
While there may be a legitimate role for the use of AI in employment litigation, given the present state of development of AI applications, and the manner applicants are commonly using them, AI is causing ongoing headaches for respondents, courts and the FWC. These issues are exacerbated by applicants either not declaring the use of AI (as sometimes required by procedural rules) or not being entirely candid in declaring the extent or nature of such use.
There are two factors that might lead to some sanity being restored.
The first is adverse costs orders being made against applicants, although the costs provisions in the Fair Work Act 2009 (Cth) (FWA) do place some limitations on the extent to which a court or the FWC can do this. Surely, however, using AI to produce purported evidence or submissions that include unchecked hallucinations, and then being less than candid about AI use when asked, is squarely in the territory of an “unreasonable act” for the purpose of section 570 (for court proceedings) or section 400A (for unfair dismissal matters in the FWC).
The second is a realisation by applicants that they are being had and the promise of AI being a substitute for an expert advisor is largely a hollow claim propagated by hype merchants. While AI might, at this point, largely be an example of the “Emperor’s New Clothes”, it is the unwitting employee applicants relying upon it being left exposed and vulnerable in a court or the FWC, with only specious documents and false hope covering their dignity.
