What happened
In Finnegan v Kemner (No 3) [2026] FCA 1402, the applicant sought to withdraw an admission he had made in a substituted reply to a defamation claim. The respondent argued that the admission was a ‘verbose ’ produced by the applicant’s AI tool, suggesting the explanation for the admission made no sense. On 21 September 2026, the Federal Court granted the applicant leave to amend the pleading, finding the applicant’s explanation plausible and noting minimal prejudice to the respondent. The court ordered the applicant to pay the respondent’s costs related to the amendment and directed the filing of an amended substituted reply. The judgment also referenced a prior insurance matter where a pleading was described as ‘AI slop’ under the court’s Generative Artificial Intelligence Practice Note (GPN‑AI).
Finnegan v Kemner (No 3) arose from an email the applicant sent on 8 April 2019, which the respondent alleged was defamatory. The respondent’s defence claimed the applicant had sent the email anonymously using an external affairs address.
The applicant initially admitted to the allegation in a substituted reply filed on 5 December 2025. On 21 September 2026, he applied for leave to amend the pleading to withdraw that admission, arguing the email did not identify him by name but was not sent anonymously.
The respondent countered that the applicant’s explanation for the admission was nonsensical and characteristic of an AI , asserting that the applicant used AI for most of his recent communications.
The Federal Court rejected the AI‑ argument, finding the applicant’s explanation plausible and ordering the applicant to pay the respondent’s costs associated with the amendment. The court emphasized that the amendment would not undermine public confidence or cause undue prejudice.
The judgment also referenced a separate insurance matter where a pleading was labelled ‘AI slop’ under the court’s Generative Artificial Intelligence Practice Note, indicating ongoing judicial attention to AI‑generated content.
Source details: thelawyermag.com ↗
Why it matters
The decision provides an early judicial assessment of how Australian courts may treat arguments that AI‑generated content is unreliable or ‘hallucinatory’. By rejecting the respondent’s AI‑ defence, the court signalled that AI‑related explanations will be scrutinised on their factual plausibility rather than dismissed outright. This sets a practical precedent for litigants who rely on for drafting pleadings, indicating that courts expect a reasonable basis for any AI‑assisted statements and will not accept AI errors as a blanket excuse. The ruling also underscores the growing relevance of procedural rules—such as the Federal Court’s GPN‑AI practice note—in shaping litigation strategy and cost allocations when AI tools are used. For legal practitioners, the case highlights the need for clear documentation of AI involvement and the potential for cost penalties if AI‑generated content is deemed insufficiently substantiated.
The ruling clarifies that AI‑related errors are not automatically excusable in court proceedings; parties must still provide a coherent factual basis for any AI‑assisted statements.
By upholding the amendment, the court affirmed that procedural flexibility can accommodate corrections without compromising the integrity of the litigation process.
The decision highlights the practical impact of the Federal Court’s GPN‑AI, which sets expectations for the quality and transparency of AI‑generated legal documents.
Cost orders against the applicant demonstrate that courts may penalise parties for inadequate handling of AI‑generated content, reinforcing the need for diligent oversight.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
Impossibility results in algorithmic fairness (e.g. Kleinberg et al., Chouldechova) show what?
What to watch next
Future Australian and Commonwealth court decisions that address AI‑generated evidence or pleadings, especially where parties invoke AI as a defence. Monitoring updates to the Federal Court’s Generative Artificial Intelligence Practice Note and any legislative proposals concerning AI use in legal proceedings will be essential. Additionally, watch for any appellate rulings that may refine the standards applied to AI‑related arguments in civil litigation.
Subsequent cases where AI is raised as a defence or explanation, particularly in defamation, contract, and tort matters.
Potential amendments to the Federal Court’s Generative Artificial Intelligence Practice Note, which could introduce stricter disclosure or verification requirements.
Legislative initiatives at the state or federal level that aim to regulate AI use in legal practice, including professional conduct rules for lawyers.
Any appellate rulings that may reinterpret the standards applied in this case, potentially affecting the broader legal landscape for AI‑assisted litigation.