On 17 June 2026, Symons J delivered judgment in the case of Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245 (Ba v Sterling Parts).  Ba v Sterling Parts was a proceeding brought by Weiman Ba (Applicant) in the second division of the Federal Circuit and Family Court of Australia (Court).  Symons J dismissed the unlawful termination case against Sterling Parts Australia Pty Ltd and Sterling Parts (Melbourne) Pty Ltd (Respondents) and warned of the risks of using generative artificial intelligence (AI) in civil litigation.

Context to the case of Ba v Sterling Parts

The underlying employment dispute arose when the self-represented Applicant filed documents in the Court on 2 February 2026 in support of his claim that the Respondents had contravened the Fair Work Act 2009 (Cth) (Fair Work Act).  The Applicant alleged that the Respondents misclassified him as an independent contractor rather than an employee.  The Applicant was a delivery driver for the Respondent from November 2021 and was terminated on 29 September 2025.

The Applicant sought compensation for unlawful termination and underpayments.  He has raised concerns regarding his working arrangements and entitlements, in connection with his classification as an independent contractor.  His claim was ultimately dismissed because he relied on an AI-generated “Evidence List” of documents that the Court inferred did not exist.

It was found that the Respondents had breached section 340(1) of the Fair Work Act, which states:

• “a person must not take adverse action against another person:
  
  
  • • because the other person:
      
      
      • • has a workplace right; or
      
      
      • • has, or has not, exercised a workplace right; or
      
      
      • proposes or proposes not to, or has at any time proposed or proposed not to, exercise a workplace right; or
  
  
  • to prevent the exercise of a workplace right by the other person”.^[1]

This subsection forms a civil penalty provision.^[2]

What was the AI-generated evidence?

In support of the Applicant’s Statement of Claim, the Applicant filed a “Schedule of Documents” (Evidence List) containing thirty-eight (38) separate entries (prefixed “EV-001” to “EV-038”).  These documents purportedly proved key indicators of an employment relationship, such as a “Vehicle & Operating Cost Policy” (EV-014), “Daily delivery schedules” (EV-028), and a “Three-Tier Management Structure” (EV-033).

The Applicants pleadings were heavily integrated with these documents, using them to substantiate his claims of operational control and economic dependence.  However, when the respondents repeatedly requested to inspect the underlying documents, the Applicant engaged in a “pattern of obstruction and prevarication”, providing only “summaries” or re-attaching the list itself rather than the actual files.^[3]

On what basis did the Court reject the evidence?

Following the Applicant being “deliberately obtuse”, the Court issued a self-executing order on 17 June 2026, requiring the Applicant to produce documents by 24 June 2026.^[4]

The Court dismissed the evidence and claim after a consideration of the following:

• • There was a non-existence of documents. In his final attempt at compliance, the Applicant produced only one (1) authentic document, a termination email.  There was a failure to produce any of the other 37 items.^[5]

• • Symons J inferred that the Evidence List was AI-generated and that the documents either did not exist, or the descriptions did not match the contents. It was raised that while AI can usefully assist with administrative tasks, it is a risk to rely on it as a legal advisor given it can prioritise user approval over factual truth, potentially leading to evidence fabrication.^[6]

• Symons J determined that the Applicant’s “blind faith” in an AI tool, relying upon it as a legal expert, is what resulted in the outcome of the claim. Because of it, the Respondents were determined to have been effectively denied procedural fairness and the ability to understand the case against them.^[7]

Pursuant to rule 5.07(a) of the Federal Circuit and Family Court of Australia (Division 2) (General Federal Law) Rules 2025 (Cth), the Court dismissed the application.^[8]  Symons J found that the Evidence List was misleading and that continued reliance on it constituted an abuse of process.^[9]

Key takeaways

The decision of the Court in Ba v Sterling Parts is important because it shows that the use of AI in litigation is not always prohibited, but any material generated with the assistance of AI must be carefully checked.  A party who files or relies upon AI-generated material remains responsible for its accuracy.  Symons J specified that AI is not a lawyer, owes no duty to the Court or to the person using it, and that reliance on the fabricated list bordered on contempt.^[10]

Links and further references

Legislation

Fair Work Act 2009 (Cth)

Federal Circuit and Family Court of Australia (Division 2) (General Federal Law) Rules 2025 (Cth)

Cases

Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245

Further information

If you are an employer and need advice disclosure and rules of evidence, contact us for a confidential and obligation‑free discussion.

Disclaimer

This article contains general commentary only.  You should not rely on the commentary as legal advice.  Specific legal advice should be obtained to ascertain how the law applies to your particular circumstances

---

[1] Fair Work Act 2009 (Cth) s 340.

[2] Fair Work Act 2009 (Cth) s 340.

[3] Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245 at [49].

[4] Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245 at [34].

[5] Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245 at [56].

[6] Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245 at [56].

[7] Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245 at [59].

[8] Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245 at [1].

[9] Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245 at [57].

[10] Ba v Sterling Parts Australia Pty Ltd [2026] FedCFamC2G 1245 at [55].

Related insights about Artificial Intelligence law