We know that AI can help lawyers improve their efficiencies to save time, but we also recognise the impact that AI is having on court proceedings and the attention that incorrect usage of AI has been attracting in the media.

To help legal practitioners navigate the rapid adoption of AI in the Australian legal industry, we have created this page to serve as a reference on the protocols for AI use in Australian court jurisdictions. Simply scroll down to your jurisdiction to learn more.

Federal Courts

The High Court of Australia

The High Court of Australia has not issued any notices or guidelines on the use of AI in court proceedings.

The High Court rules have not yet been updated to specifically address the use of Artificial Intelligence in court proceedings. While no AI legal regulations exist at this level, the High Court is monitoring developments in technology and its potential impact on the justice system.

The Federal Court of Australia

The Chief Justice released a media statement in March 2025 expressing that the court expects any practitioners and self-represented litigants using Generative AI to do so responsibly, consistent with existing obligations, and disclose such use when required by a Judge or Registrar.

Following this and after a consultation process in 2025, the Chief Justice released a Practice Note on the Use of Generative Artificial Intelligence (GPN-AI) in April 2026. It applies to all persons who appear before or file documents with the Federal Court, including litigants (including self-represented), witnesses, and third parties producing documents under subpoena or other orders. While the Court recognises the potential benefits of Generative AI for efficiencies, it has set out clear expectations on responsible use.

Expectations:

• If using AI, you must have a basic understanding of its capabilities, limitations and risks.
• Usage must not adversely affect the administration of justice.
• Users remain bound by their existing legal and professional responsibilities.
• If the Court requires it, a person must disclose to the Court if (and how) Generative AI has been used in a proceeding. Users should be in a position to inform the Court what tool was used, how it was used, and for what purpose.

Regarding pleadings, written submissions, and other filed documents

Where Generative AI has been used to prepare documents, the responsible person must confirm that:

• Facts stated in pleadings are based on what the party reasonably considers can be proved, and claims for relief are based on proper legal principles
• Legal authorities cited in submissions exist and support the proposition stated
• Evidence cited exists, is or will be in the materials before the Court, and is reasonably likely to be admissible
• Statements about what the evidence proves are findings reasonably open for the Court to make
• Chronologies are accurate
• Lists of documents conform with the Federal Court Rules 2011 (Cth): rr 20.17 (form of list), 20.14 (standard discovery) or 20.15 (non-standard discovery)

Regarding affidavits, expert reports, and other evidentiary materials

Witnesses, parties and their lawyers must be conscious of their obligations when using AI to assist in preparing evidence. Affidavits and witness statements must reflect the deponent’s own recollection, knowledge or experience, and expert reports must contain the expert’s own opinion and process of reasoning.

Disclosure of Generative AI use is required where AI tools were used:

• To summarise or analyse information upon which a witness relies to make a statement of fact or express an opinion
• To create images, video, sound recordings or other multimedia presented to the Court (which should be clearly identified as having been produced using AI)
• In any other manner that might reasonably be expected to affect the admissibility of the evidence, or the use the Court makes of it

Disclosure should occur in the body of the document at the start of the document, and should concisely identify where in the document Generative AI has been used and how.

Using AI with confidential, suppressed, or private information

Users must carefully consider, before inputting information into an AI tool, whether the information is:

• The subject of Court orders as to confidentiality, suppression or non-publication
• Privileged
• Subject to the implied obligation not to use it for a purpose other than the proceedings in which it was obtained
• Otherwise confidential or private, without the consent of the person to whom it is confidential or private

The Practice Note warns that information provided to a generally accessible AI tool may become available to other people, and that users may not know where it is stored, how it is used, or who can access it. The risks of inadvertent disclosure may be lower for “closed” or “controlled” AI tools, but users of those tools are still expected to rigorously preserve the integrity and confidentiality of information obtained through compulsory processes in proceedings. There may be serious consequences for entering protected information into Generative AI tools, even if sharing it was not intended.

Federal Circuit and Family Court of Australia

In June 2026, the FCFCOA released a Practice Direction that applies to all proceedings and is intended to promote responsible AI practices that maintain confidence in the Courts. It applies to all litigants (including self-represented), lawyers, experts, and other users the rules for using AI in the Federal Circuit and Family Court of Australia.

The Practice Direction sets clear expectations for how court users may use AI, reinforcing that professional duties and obligations continue to apply. Court users still remain responsible for the accuracy and authenticity of any material filed or relied on. Any files or documents prepared with the help of GenAI must be critically assessed, fact-checked and verified (including citations to legislation and authorities). It also emphasises confidentiality, privacy and safety risks, and strict limits on entering suppressed, subpoenaed, non-public, or otherwise protected information into GenAI tools unless the user is satisfied it will remain within a closed environment and will not be used to train models. The Court may respond to inappropriate use of AI by disregarding material, making costs orders, or taking other action.

Although it isn’t a requirement for a court users to disclose the use of AI, if you do use an AI tool in the preparation of a document, you may be asked to. If you are, it is expected you know:

• What tool was used
• How the output was checked or supervised
• how the principles in the Practice Document were observed.

The Court also distinguishes between:

• "Closed AI" systems: when an AI tool contains controls for the use of information, and protections from information being made public or being used to train the AI tool, such as Smokeball’s AI features.
• "Public AI" systems: an AI that is publicly available via the internet (e.g. ChatGPT, Claude, and Gemini) and does not have controls as to the use or further dissemination of information. Public AI tools are trained on broad, often public datasets, most commonly for general purpose use. The Practice Direction specifically warns against entering confidential, protected or sensitive information into public AI systems.

‍

New South Wales

Updates to the Uniform Civil Procedure Rules 2005

The Uniform Civil Procedure Rules 2005 (NSW) were recently amended (Amendment No.104 of 2025) to help regulate the use of generative AI:

• Rule 31.4 (3A-3C) states that generative AI cannot be used to write or alter a witness statement and they must now state AI was not used. It also cannot be used to generate the content of an annexure or exhibit to a witness statement.
• Rule 35.3B states that generative AI cannot be used to write or alter affidavit content. It also requires every affidavit to state that generative AI was not used to write it.

These updates reflect the practice note released by the Supreme Court.

Supreme Court of NSW

The regulations around the use of generative AI in NSW are outlined in the Supreme Court’s *Practice Note SC GEN 23 — Use of Generative Artificial Intelligence (Gen AI)*, effective from 3 February 2025.

The allowed uses of Generative AI, subject to safeguards

• Generating chronologies, indexes, and witness lists
• Preparing briefs or drafting Crown Case Statements
• Summarising or reviewing documents and transcripts
• Assisting with written submissions, provided the author personally verifies all citations, authorities, legislation, and references to evidence for existence, accuracy, and relevance, and does not rely on Gen AI for verification.

Prohibitions and strict limits to the use of Generative AI

• Do not enter suppressed or non-public information, subpoenaed material, or information covered by the Harman undertaking into Gen AI unless all of the following are satisfied:
  • The platform is controlled, under confidentiality restrictions, and will not use the data to train models
  • The data is used only for the proceeding
  • The data is not used to train any Gen AI or large language model
• In the case of affidavits and witness materials:
  • Gen AI must not generate the content of affidavits, witness statements, character references, or other evidentiary material
  • Gen AI must not alter, embellish, strengthen, dilute, or rephrase a witness’s evidence
  • Affidavits, witness statements, and character references must disclose that Gen AI was not used to generate their content. Annexures/exhibits prepared for the purposes of the evidence also require disclosure, unless leave is granted. Using Gen AI for annexures or exhibits requires leave with detailed particulars of the tool and proposed use
• In the case of written submissions:
  • Authors must personally verify all citations, legal authorities, legislation, and references to evidence
  • Verification must not be done solely by Gen AI
  • Use of Gen AI does not lessen professional or ethical obligations
• In the case of expert reports:
  • Gen AI must not draft or prepare any part of an expert report without prior leave of the Court
  • If leave is granted, the expert must disclose which parts used Gen AI, identify the program and version, keep records of prompts and settings unless dispensed with, and annex any applicable professional codes
  • Special timing for leave applies to certain professional negligence and work injury damages matters (seek leave at the first directions hearing). Reports prepared between issue and commencement must identify any reliance on Gen AI.

Key cautions and risks highlighted by the Court

• Hallucinations, false citations, outdated or irrelevant data, bias, confidentiality and privilege loss, unintended training on inputs, and potential copyright issues are all risks associated with Gen AI use. Users must understand tool limits and ensure privacy and data security settings are appropriate.

District Court of NSW

The District Court General Practice Note 2 - Generative AI Practice Note and Judicial Guidelines formally adopts the Supreme Court’s rules above, effective 3 Feb 2025.

NSW Land & Environment Court

The Practice Note issued by the Land & Environment Court - Use of Generative Artificial Intelligence (effective 12 Feb 2025) - is identical to the Supreme Court’s.

NSW Civil and Administrative Tribunal

NCAT Procedural Direction No. 7 (effective April 2025) applies the same rules as the Supreme Court in tribunal proceedings.

The Law Society of New South Wales

In December 2024, a joint statement was released by the Law Society of NSW, the Legal Practice Board of Western Australia, and the Victoria Legal Services on the use of AI in Australian legal practice. The statement advises that legal professionals must exercise utmost caution when using AI tools in their practice, ensuring strict adherence to ethical standards under Uniform Law and Solicitors’ Conduct Rules. Key suggestions mirror the Supreme Courts, such as:

• Maintaining client confidentiality by avoiding input of sensitive data into public AI platforms,
• Personally verifying all AI-assisted documents to prevent inaccurate or misleading information
• Restricting AI usage to low-risk, easily verifiable tasks
• Implementing clear policies and active supervision
• Transparently recording and disclosing AI use to clients and courts

‍

Victoria

There has been no change in legislation on the use of AI in court proceedings in Victoria, though guidelines have been issued.

Supreme Court of Victoria

In May 2024, the Victorian Supreme Court published guidelines for litigants on the responsible use of AI.

The guidelines emphasis that users must understand AI limitations, maintain confidentiality, and be transparent about AI use. Litigants are reminded to uphold their duty of candour, while self-represented parties should disclose AI tools used. Special caution is advised for affidavits and witness materials, which must reflect the person's own knowledge. Specialised legal AI tools (such as Archie AI) are recommended over generic chatbots like ChatGPT. We have broken them down below.

Permitted and encouraged uses of Artificial Intelligence

• Use AI that searches within a closed corpus to improve efficiency, such as Technology Assisted Review (TAR) for large-scale discovery, with cooperation between parties and court management.
• Prefer to use specialised, legally-focused AI tools and databases over general-purpose chatbots, which are not secure and are not trained for the Australian legal industry.
• Self-represented litigants and witnesses who use generative AI are encouraged to include a statement identifying which tools they used in filed documents or reports.

Disclosure and responsibilities of litigants

• Parties and practitioners should disclose AI assistance to each other, and where appropriate to the Court, so others can assess provenance and weight
• Signing or filing a document represents that it is accurate and complete - there is no defence for errors or omissions caused by AI assistance.
• Use of AI is still subject to existing obligations, including candour to the Court and the Civil Procedure Act 2010 requirements that documents and submissions have a proper basis.

Cautions and limits on AI usage

• Litigants should understand how the AI tool works and its limitations
• Be alert to privacy and confidentiality risks when providing information to external AI services
• Treat general-purpose large language models (LLMs, such as ChatGPT) as non-reasoning, non-research tools. Their outputs must be checked and are more likely to be unsuitable for current litigation.
• Any AI-generated text should be carefully reviewed to ensure it is not:
  • Out of date
  • Incomplete
  • Inaccurate or incorrect
  • Inapplicable to Victorian or Australian law
  • Biased

Prohibitions and strong discouragements

• Do not mislead the Court or other participants about the nature of work undertaken or the content produced by AI.
• Exercise particular caution with affidavits, witness statements, and expert materials. Evidence must reflect the witness’s own knowledge and words, and experts must comply with the Expert Witness Code of Conduct.
• The Court does not use AI to make decisions or draft reasons, because AI does not engage in case-specific reasoning.

Victorian County Court

In July 2024, the Victorian County Court issued the practice note Guidelines for Litigants: Responsible use of artificial intelligence in litigation ****that cover the same principles as above in the County Court.

Law Institute of Victoria

The LIV provides an AI hub of resources, and the Victorian Legal Services Board contributed to the joint statement on AI use in December 2024.

‍

Queensland

No Queensland laws directly govern AI in court proceedings.

In May 2024, a set of guidelines were published for self-represented litigants that apply to all QLD courts. They explain that AI chatbots can help with organising facts or drafting outlines, but they are limited by their training, the prompts users write, and their lack of context to your situation.

In September 2025, these guidelines were amended to include a caution for using AI to help prepare an affidavit or witness statement, stating that any documents must be sworn or finalised to accurately reflect a person’s own knowledge and words. The same update added a warning to self-represented litigants that relying on fake or inaccurate AI-generated information in court could result in court delays and costs to them.

A set of guidelines for judicial officers that was also developed in May 2024, were made available publicly.

Queensland Law Society

In June 2024, the QLS issued Guidance Statement No. 37 Artificial Intelligence in Legal Practice, reminding solicitors to apply ethical duties when using AI and stressing that lawyers must exercise independent judgment, maintain confidentiality, ensure competence, and verify AI output.

The Queensland Bar has also advised barristers to follow their professional obligations when using AI.

Otherwise, QLD lawyers remain bound by the Uniform Law and ASCR/BR duties on honesty, competence and confidentiality, as emphasized in the joint national statement.

‍

Western Australia

As of mid-2025, there are no formal rules or practice notes released on the use of AI.

Supreme Court of WA

In March 2025, the Supreme Court published a consultation note on Generative AI use that asked for input on a practice direction regarding AI. The consultation is intended to shape WA-specific guidance for lawyers.

Legal Practice Board of Western Australia

The Legal Practice Board of WA contributed to the joint statement on AI use in December 2024.

‍

South Australia

South Australian courts have not issued AI-specific protocols yet, however in May 2025, the Chief Justice of SA sent a survey to the profession on generative AI use, the results of which are intended to inform a subcommittee tasked with framing potential rules.

Professionally, the Law Society of SA (LSSA) are actively engaging on AI issues. In June 2025, they reported submitting feedback on the Federal Court’s AI proposal and they continue to advocated mandatory disclosure of AI use.

‍

Tasmania

Tasmania has no AI-specific court laws or amendments as of mid-2025, nor have any Tasmanian courts issued practice directions or guidelines on AI use.

‍

Australian Capital Territory

The ACT has no AI-specific court laws or amendments as of mid-2025, nor have any courts issued practice directions or guidelines on AI use.

‍

Northern Territory

The Northern Territory has no AI-specific court laws or amendments as of mid-2025, nor have any courts issued practice directions or guidelines on AI use.

‍

How can you use AI responsibly in court?

• Understand AI limitations: Be aware of what AI tools can and cannot do, and use them appropriately for their intended purpose.
• Restrict AI to low-risk tasks: Use AI primarily for administrative or preparatory work rather than legal reasoning.
• Maintain confidentiality: Avoid inputting sensitive client data into public AI platforms like ChatGPT.
• Verify all outputs: Personally check all AI-generated content for accuracy, currency, and relevance to Australian law.
• Use legal-specific AI: Prefer specialised legal AI tools like Archie AI over general-purpose chatbots like ChatGPT.
• Practice transparency: Disclose AI use to clients, opposing parties, and the court when appropriate.
• Exercise caution with evidence: Ensure affidavits and witness statements reflect the person's own knowledge and words.
• Follow court-specific guidelines: Adhere to any practice directions or protocols issued by the relevant jurisdiction.
• Maintain professional responsibility: Remember that you remain responsible for all documents filed, even if AI-assisted.
• Document AI usage: Keep records of how and where AI tools were used in preparation of court documents.

‍

AI isn’t for replacing lawyers, but supporting them.

As AI continues to transform legal practice in Australia, responsible use requires balancing innovation with ethical obligations and court protocols. By following these guidelines, legal professionals can harness AI's benefits while maintaining the integrity of our justice system and upholding their duty to the court.

‍

Last Updated: 19 September 2025.