Artificial intelligence (AI) tools are now widely used in personal, business and legal settings. They can assist with a range of legal tasks when used appropriately.

However, there are important risks and limitations that you should be aware of if you are considering using AI tools in connection with legal matters.


AI Is Not a Substitute for Understanding


As our client, it is very important to us that you understand the legal work we are doing for you, the advice we provide, and the decisions that need to be made throughout your matter.


We want you to make informed decisions about your legal situation, so we can act for you and take instructions from you, not from an AI system you may use. If AI has helped prepare your instructions, responses, explanations, or documents, it remains important that you read that material carefully and ensure it accurately reflects your knowledge, intentions, and position.


Impact on Legal Costs


Providing AI instructions, documents or summaries to us can increase the amount of work required, and the cost of our legal services to you.


For example, additional time and work may be needed where:


  1. You provide large volumes of AI-generated material which must be reviewed;
  2. We are instructed by you to answer questions or do work that an AI recommended, but which was optional; or
  3. AI-generated content contains inaccuracies, unsupported assertions, or fabricated information, and needs to be thoroughly checked.


The more AI-generated material that requires checking and verification, the greater the potential impact on legal costs.


AI Can Be Wrong or Incomplete


AI systems do not independently verify facts and can generate information that is inaccurate, incomplete, misleading, outdated, or entirely incorrect.


AI-generated content should not be assumed to be accurate simply because it is presented confidently or appears professional. Important details, qualifications, dates, documents, legal authorities, or factual matters may be omitted or misstated.


For this reason, all information and documents provided to us should be carefully reviewed before being relied upon. We will undertake our own assessment of material provided, but clients remain responsible for ensuring that instructions given to us are accurate and complete to the best of their knowledge.


We frequently notice AI making mistakes - it sometimes relies on out-of-date laws, invents cases that don’t exist, or justifies conclusions by citing legal authorities that aren’t relevant.


AI May Tell You What You Want to Hear


AI tools are prediction machines that generate text or images without actually understanding. They typically generate responses that are very confident, and appear helpful, persuasive, and responsive to the user's request. But an answer is only as good as the prompt, and the AI tool. There may be information that is relevant to your legal position that you don’t know should be considered, so didn’t provide to the AI. It can be very difficult to write a prompt that isn’t biased toward the outcome that you want, and AI can usually pick up on that.


AI may produce answers that are more optimistic than the evidence supports. In a legal context, this can create an unrealistic impression of the strength of a claim, defence, negotiation position, or likely outcome.


We treat AI-generated advice or analysis with caution until it is verified. A legal assessment requires consideration of all relevant facts, evidence, legal authorities, risks, and potential weaknesses, not simply the information or assumptions that have been presented to the AI system.


Protecting Legal Privilege and Confidentiality


In a dispute, some communications or documents may be legally privileged and don’t need to be disclosed. This can include what you tell us, the advice we give you, and documents you have prepared in anticipation of a court case.


Legal privilege can be lost if privileged material is disclosed to third parties or treated as if it is not confidential. This could mean that if you are involved in a dispute that goes to Court, you might be required to disclose all documents shared with AI to the other party.


Many AI platforms process information on servers operated by external providers (often overseas) and may use the information provided by users for training, analysis, or other purposes. Entering communications with your lawyer, legal advice, draft affidavits, witness statements, settlement discussions, or other privileged material, into a publicly available AI system may create a risk that privilege could be compromised.


Before entering any information relating to your legal matter into an AI system, especially a public system that is free to use, you should know whether confidentiality is assured and whether that AI use could cause the information to lose the protection of legal privilege.


If in Doubt


Additional information has been released by the Ministry of Justice and can be found by searching for: Guidelines for use of generative artificial intelligence in Courts and Tribunals - Non-lawyers.


If you are considering using AI in connection with your legal matter, or have already done so, please let us know. Early disclosure helps us protect confidentiality, assess any risks, and provide the most efficient and effective advice possible.


If you have any questions, please raise these with the staff-member assisting you with your matter.