A protocol for accuracy, authority and accountability

Generative AI is rapidly transforming legal practice, offering powerful support for drafting, research, and document analysis, but it also introduces serious professional risks. Courts in countries like the United States, United Kingdom, and Canada have already sanctioned lawyers for relying on AI-generated material containing fabricated case law, highlighting the need for strict verification processes. At the same time, the use of AI raises critical privacy and confidentiality concerns under frameworks such as New Zealand’s Privacy Act 2020 and professional conduct rules, especially when handling sensitive client information.

Authored by Kalev Crossland and Kathryn Dalziel, this protocol provides practical guidance to help lawyers use AI responsibly. It includes safeguards to prevent false authorities, explains the consequences of misuse, and outlines how to manage privacy risks across both cloud-based and emerging local AI tools. It also examines evolving regulatory responses and offers forward-looking guidance tailored to legal practice in New Zealand.