AI rules for lawyers, by country · New Zealand · Asia Pacific
New Zealand: three audiences, three documents
Guidelines for Use of Generative Artificial Intelligence in Courts and Tribunals · Courts of New Zealand · December 2023
Page verified 5 August 2026. We could not open the guidelines from this session, so everything below is recorded from contemporaneous reporting and official statements rather than quoted from the document. We say so rather than imply we read it.
New Zealand did something no other jurisdiction has done: it wrote three separate documents for three separate audiences, including one for people who have no lawyer at all. That third document is the only one of its kind anywhere on this record.
The three sets
- LawyersExisting professional obligations apply to these technologies. Exercise caution with generative AI chatbots because of their risks and limitations.
- JudgesA separate set for judges and judicial officers, the same architecture England and Wales and Hong Kong later adopted.
- Non-lawyersA set written for self-represented litigants and anyone else appearing without counsel.
All three issued in December 2023, which makes New Zealand one of the earliest movers anywhere, ahead of the England and Wales judicial guidance of the same month and well ahead of the practice directions in Qatar and New South Wales.
Why the third document matters
Most instruments on this record assume a professional on the other side of the duty. The population actually most exposed to AI hallucination is the one that cannot afford a lawyer, does not know that free case-law databases exist, and has no way to tell a real citation from an invented one.
Harber v HMRC is that population in a courtroom. The Tribunal found as a fact that the appellant did not know the cases were fabricated and did not know how to check them. No sanction followed, because none was appropriate. New Zealand had already written the document that would have helped her, two years earlier and in a different hemisphere.
Connecticut’s Practice Book section 4-9 and New South Wales’s SC Gen 23 both bind unrepresented parties. New Zealand is the jurisdiction that also told them how.
What to do about it
Seven days free, then $349 a month. Or run it across the estate.
Existing professional obligations applied to a new technology is the same conclusion every jurisdiction has reached. What differs is only what you can show afterwards. The same engine either way: a signed record made on your own machine, content-free, and checkable free by anyone, forever.
Start the free weekOr Regulayer for enterprise →
See also the attorney certification, the free verifier, and every jurisdiction we track.
Sources
- Courts of New Zealand, Guidelines for use of generative artificial intelligence in Courts and Tribunals
- Not opened from this session. Recorded from the Courts of New Zealand listing and contemporaneous commentary. We carry no quotation.
Related
- AI rules for lawyers, country by country
- Harber v HMRC, the litigant in person
- Australia
- New Zealand · Privacy Act 2020
Checked 5 August 2026. Not read at source. Information, not legal advice.
Common questions
Does New Zealand have rules on AI in court?
It has guidelines rather than a binding practice note. In December 2023 the Courts of New Zealand issued three sets of Guidelines for Use of Generative Artificial Intelligence in Courts and Tribunals: one for lawyers, one for judges and judicial officers, and one for non-lawyers.
What do the New Zealand guidelines tell lawyers?
That their existing professional obligations apply to the use of these technologies, and that they should exercise caution when using generative AI chatbots because of the risks and limitations of those tools. The guidelines create no new duty; they apply the existing ones.
Why does New Zealand have a separate guideline for non-lawyers?
Because self-represented litigants are a substantial part of the population that uses AI for legal work, and they are the group least equipped to check its output. Harber v HMRC in the United Kingdom is what that looks like when it goes wrong: a litigant in person cited nine fabricated tribunal decisions and did not know how to check them.
Has the New Zealand Supreme Court commented on AI use?
It has been reported as issuing a warning about AI use in litigation proceedings. We have not read that decision and do not describe it further.
Information, not legal advice, and not advice on the law of any jurisdiction. Every entry is verified against the issuing body’s own document; where a source is reporting rather than the document, we say so.