There are many ways that AI can help in preparing materials for a Court proceeding, but there have been a number of cases where using AI has not worked out well and can have serious consequences.
AI hallucinations being included in submissions has caused serious problems for litigants. The clearest warning statement has come from the Supreme Court in Jones v Family Court at Whangārei.[1] The Supreme Court explained the responsibility that anyone in Court – both lawyers and non-lawyers – has have to only use accurate information in Court:
“Persons filing submissions in Court must ensure all authorities referred to are genuine and correctly cited. The current guideline for non-lawyers appearing in Court proceedings reads relevantly as follows:
You are responsible for ensuring that all information you provide to the court/tribunal is accurate. You must check the accuracy of any information you get from a [generative AI] chatbot before using that information in court/tribunal proceedings.
Reliance on false citations, including the unverified outputs of AI applications, may in serious cases amount to obstruction of justice or contempt of Court.”
In several cases, courts have reminded parties about the need to be accurate and check information provided by generative AI.[2]
As these issues become more well-known, Chief Justice Helen Winkelmann said at a New Zealand Law Society event that “The courts have shown a high degree of tolerance, but that tolerance has its limits.”
The Court can award costs where using AI wastes time or increases the costs of the other parties or the Court. In O’Driscoll v Rehab Co Mobile Ltd,[3] the Employment Relations Authority awarded costs against a self-represented litigant, saying that referring to cases that did not exist in oral submissions and final written submissions increased the other parties’ costs.
Another Employment Relations Authority decision looked at an employee uploading information into an AI platform.[4] By doing that, the employee was breaching their contractual obligations regarding confidential information and the Authority issued a compliance order.
The Tenancy Tribunal is drafting guidelines for the use of AI, in an effort to make sure that parties use AI in a proportionate and appropriate way and not in a way which could increase costs or delays.
In cases overseas, parties have uploaded legal advice into AI platforms which has been found to be waiving lawyer-client privilege. It’s critically important that clients take care not to accidentally waive privilege as it means that documents that would otherwise be protected have to be given to the other party.
Some of the early themes coming from AI use in the Court are the risks of uploading privileged or confidential information into AI platforms, the ability of the Court to impose consequences for wrongly using AI, and the importance of checking information so that only accurate information is used in Court.
[1] Jones v Family Court at Whangārei [2026] NZSC 1
[2] Wikeley v Kea Investments Ltd [2024] NZCA 609 at note 187; LMN v STC [2025] NZEmpC 46
[3] O’Driscoll v Rehab Co Mobile Ltd [2026] NZERA 106
[4] QTR v BXD [2025] NZERA 716
