Supreme Court of Victoria Practice Note SC Gen 25, commencing 14 May 2026, for every court user in that Court. Recorded 10 Oct 2026: Practice Note for Court users and Guidelines for judicial officers on the use of AI; Practice Note PNVCAT11 – Use of generative artificial intelligence.
Supreme Court of Victoria Practice Note SC Gen 25 commenced on 14 May 2026 and applies to all court users, including lawyers, parties and witnesses. It says court users may use AI in court documents if they comply with the note. Compliance includes meaningful human verification. One AI tool cannot be used to confirm another.
Court users may use AI in court documents if they comply with the note. Compliance includes meaningful human verification. One AI tool cannot be used to confirm another.
A County Court matter does not become a Supreme Court matter because SC Gen 25 is the better-known note. Open the County Court’s guidelines and follow those.
County Court of Victoria — Issued 5 June 2026. It supersedes the County Court’s 3 July 2024 litigant guidelines. It does not require a disclosure statement. Court users must be prepared to identify portions produced with AI and explain how the output was verified. The note uses its own paragraph numbering. It is not a copy of the Supreme Court note’s paragraph numbers.
Supreme Court of Victoria — These are guidelines for judges, associate judges and judicial registrars of the Supreme Court of Victoria, including reserves. They are not a practice note binding the parties. The PDF is dated 14 May 2026. They are informed by eight principles in the Victorian Law Reform Commission report of October 2025. Judicial officers must not use generative artificial intelligence for judicial decision-making. They exercise their own judgment and give reasons. The guidelines do not prevent supportive uses by officers or staff. Permissible supportive uses are organising and locating case materials, summaries and chronologies, an aid to legal research, and proof-reading. Those uses are not a substitute for reading or listening to the evidence and submissions, or for fact-finding. For case-related matters, only Court-approved tools may be used, with confidentiality controls. Officers should direct staff on use and disclosure. The guidelines will be reviewed. This is guidance for judicial officers.
Coroners Court of Victoria — Coroners Court of Victoria AI pilot to improve workflows, applying Victorian Government and Court Services Victoria guidance.
Supreme Court of Victoria — Use of AI must not indirectly mislead the Court or others about the nature of work or content; AI-assisted legal tasks remain subject to practitioners' duties including candour.
Victorian Civil and Administrative Tribunal — Describes acceptable use of Gen AI tools in all VCAT proceedings, principles for use and when use should be disclosed.
Victorian Civil and Administrative Tribunal — VCAT hearing-day guidance on recording and devices, stating that AI meeting assistants cannot be used because they automatically record hearings.
Victorian Civil and Administrative Tribunal — VCAT expert-evidence practice note (effective 1 February 2026) that requires experts to be made aware of PNVCAT11 on Gen AI and to include a Gen AI disclosure in expert reports where relevant.
Victorian Law Reform Commission — VLRC final report, tabled on 3 February 2026, with 30 recommendations. It proposes a principles-based approach, eight principles for safe AI use, and guidelines for court users, judicial officers and staff, plus governance and training. The recommendations are not binding until a court or tribunal adopts them. Earlier batches hold the consultation paper, not this report.
Mental Health Tribunal (Victoria) — No AI guidance found as of 10 Oct 2026 (homepage and News & Events checked). VCAT handles guardianship and is covered by PNVCAT11.
Victorian Legal Services Board + Commissioner (with Uniform Law regulators) — Dated 6 December 2024. Joint statement of the Law Society of New South Wales, the Legal Practice Board of Western Australia and the Victorian Legal Services Board and Commissioner. It applies existing Legal Profession Uniform Law duties; it does not create a new rule. Lawyers cannot safely enter confidential, sensitive or privileged client information into public AI chatbots or copilots such as ChatGPT, or any other public tool. Commercial tools used with client information require a careful review of the contract so the information will be kept secure (ASCR r 9.1; Barristers' Rules r 114). AI chatbots, copilots and other tools based on large language models cannot reason, understand or advise. The lawyer must exercise their own forensic judgment and must not treat the output as a substitute for their own assessment (ASCR r 4.1.4; Barristers' Rules rr 3(b), 4(e) and 42). No tool based on current large language models can be free of hallucinations. A lawyer who uses AI to prepare documents must be qualified to verify them and must actually ensure they are accurate and not likely to mislead the client, the court or another party (ASCR rr 4.1.2 and 4.1.3; Barristers' Rules rr 4(c)-(d), 8(a) and 35). Time and work items billed must accurately represent work done by the practice. AI use should not increase the client's costs above traditional methods, for example by extra time spent verifying or correcting output (Uniform Law ss 172-173; ASCR r 12.2). The statement also says lawyers using AI should consider risk-based policies on which tools may be used, by whom, for what, and with what information, including supervision of junior and support staff and verification before documents are settled; should limit use to lower-risk tasks that are easier to verify, with the examples of drafting a polite email or suggesting how to structure an argument; and should not use it for higher-risk tasks given as examples of translating advice into another language, analysing an unfamiliar legal concept, or executive decision-making. They should record and, where necessary or appropriate, disclose to the client, the court and fellow practitioners when and how AI was used, and how that use is reflected in costs if the client asks. The regulators say they will review the guidance as use evolves.
Victorian Legal Services Board and Commissioner — Victorian Legal Services Board and Commissioner publishes “New research examines AI use in the Victorian legal profession | VLSB+C”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board and Commissioner — Victorian Legal Services Board and Commissioner publishes “Generative AI Use in the Legal Profession – Research Brief”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board and Commissioner — Victorian Legal Services Board and Commissioner publishes “Tapping the potential of AI in legal practice: daiquiris, anyone? | VLSB+C”. New research finds that while generative AI is gaining ground in Victorian legal practice, its benefits remain narrowly conceived and unevenly distributed. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board + Commissioner — VLSB+C's 2026 risk outlook names AI as a professional competency risk, encouraging responsible use while warning that unverified GenAI output breaches professional obligations and may lead to disciplinary action.
Office of the Victorian Information Commissioner — Office of the Victorian Information Commissioner publishes “Artificial Intelligence and Privacy – Issues and Challenges”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board + Commissioner — VLSB+C's 2024 risk outlook identifying improper use of AI as a key risk and listing the regulator's expectations of lawyers who use AI.
Office of the Victorian Information Commissioner — Office of the Victorian Information Commissioner publishes “Artificial Intelligence – Understanding Privacy Obligations”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board + Commissioner — Early VLSB+C commentary (28 September 2023) on ChatGPT and other generative AI platforms: how they work, accuracy limits and the ethical rules that apply when lawyers use them.
Office of the Victorian Information Commissioner — Office of the Victorian Information Commissioner publishes “Public consultation on Artificial Intelligence privacy guidance”. Checked 2026-10-10. Read the linked page before relying on it.
Office of the Victorian Information Commissioner — Office of the Victorian Information Commissioner publishes “Use of enterprise Generative AI tools in the Victorian public sector”. Checked 2026-10-10. Read the linked page before relying on it.
Law Institute of Victoria — Law Institute of Victoria publishes “Artificial Intelligence Hub - Law Institute of Victoria”. Share your feedback to help us enhance our support on AI topics. Checked 2026-10-10. Read the linked page before relying on it.
Law Institute of Victoria — Approved by the CEO of the Law Institute of Victoria on 13 August 2025, after consultation with the VLSB+C and the LPLC. The guideline says these are guidelines only and do not have the force of law. They apply to practitioners and support staff in a law practice, and to other organisations that deliver legal services. Practitioners who use AI should: not enter confidential or legally privileged information into an open-source or commercial AI system without client consent; check all outputs for accuracy and suitability; check that cases exist and are relevant; tell the client before AI is used; be able to tell the opponent and the court, if asked, that AI was used; read the privacy policy and terms before first use and review them regularly; charge fairly; and retain what is needed for conflict checks after the retainer ends. The practitioner remains accountable for documents tendered and cannot delegate that to an AI system, staff or the client. The guideline says practitioners should be cautious until they have received appropriate training. Not every use must be disclosed if the text has been thoroughly checked, but the practitioner must know when a document was drafted with AI and be able to tell the court its origin. Tell the client before AI is used in the matter. Do not input confidential or privileged information except into a closed system whose claim to be closed has been carefully examined, and only after weighing the risk. Seek the client's consent before using confidential information with an AI model. An information barrier can be undermined if a closed system retains one client's information and later uses it for another. Principals are responsible for reasonable supervision, including of support staff who can access AI (ASCR r 37.1). A time-based bill cannot charge two hours for a document generated with AI in 20 minutes, including time spent instructing the AI and reviewing the result. Costs related to AI use are to appear in the costs disclosure and estimate.
Law Institute of Victoria — Same Law Institute of Victoria guidance note as the catalogue row for this PDF without the tracking parameter. Approved by the CEO on 13 August 2025, after consultation with the VLSB+C and the LPLC. The guideline says these are guidelines only and do not have the force of law. The 6 December 2024 date in the document is a footnote citation of the Uniform Law statement, not the approval date of this note.
Law Institute of Victoria — Law Institute of Victoria publishes “Artificial Intelligence Hub - Law Institute of Victoria”. Share your feedback to help us enhance our support on AI topics. Checked 2026-10-10. Read the linked page before relying on it.
Law Institute of Victoria — Law Institute of Victoria publishes “Artificial Intelligence and Lawyers' Ethical Duties”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Bar — Victorian Bar publishes “Artificial intelligence & large language models - Victorian Bar”. Checked 2026-10-10. Read the linked page before relying on it.
Victoria Legal Aid — Victoria Legal Aid publishes “Upgrading our online chat to provide a better service to help seekers”. The questions our digital assistant asks you are the same for every chat. The tool does not use artificial intelligence, think, or learn on its own. Checked 2026-10-10. Read the linked page before relying on it.
Victoria Legal Aid — Victoria Legal Aid publishes “Making community legal information | Victoria Legal Aid”. Artificial intelligence (AI) can support community legal information work. We find it useful for some tasks, while recognising that it also has risks and limitations. Checked 2026-10-10. Read the linked page before relying on it.
Victoria Legal Aid — Victoria Legal Aid publishes “Protecting privacy and data – obligations for panel practitioners”. The way legal services are delivered is rapidly evolving, shaped by digital communication, remote work, and the growing use of artificial intelligence. Checked 2026-10-10. Read the linked page before relying on it.
Law Institute of Victoria — LIJ report on the Victorian Justice Leaders AI Forum convened by the LIV with the Victorian Attorney-General, and the published summary of enablers and actions for AI adoption in the justice sector.
Office of the Victorian Information Commissioner — Privacy and Data Protection Act 2014 (IPPs; Victorian Protective Data Security Standards); Health Records Act 2001. FOI: Freedom of Information Act 1982 (Vic). OVIC guidance covers AI privacy obligations and enterprise GenAI tools in the Victorian public sector. VLA tells panel practitioners how to protect privacy and data. Checked 2026-10-10.
Office of Public Prosecutions Victoria — No AI statement or guidance found on the OPP Victoria homepage, and search found only vendor case studies (Appian) on 10 Oct 2026.
Issued 5 June 2026. It supersedes the County Court’s 3 July 2024 litigant guidelines. It does not require a disclosure statement. Court users must be prepared to identify portions produced with AI and explain how the output was verified. The note uses its own paragraph numbering. It is not a copy of the Supreme Court note’s paragraph numbers.
Court users must be prepared to identify the specific portions of Court documents which were produced using AI and be able to explain how the output was verified.
The guideline PDF is dated 14 May 2026. · Effective 2026-05-14
These are guidelines for judges, associate judges and judicial registrars of the Supreme Court of Victoria, including reserves. They are not a practice note binding the parties. The PDF is dated 14 May 2026. They are informed by eight principles in the Victorian Law Reform Commission report of October 2025. Judicial officers must not use generative artificial intelligence for judicial decision-making. They exercise their own judgment and give reasons. The guidelines do not prevent supportive uses by officers or staff. Permissible supportive uses are organising and locating case materials, summaries and chronologies, an aid to legal research, and proof-reading. Those uses are not a substitute for reading or listening to the evidence and submissions, or for fact-finding. For case-related matters, only Court-approved tools may be used, with confidentiality controls. Officers should direct staff on use and disclosure. The guidelines will be reviewed. This is guidance for judicial officers.
Guidance for judicial officers, not a practice note binding the parties.
Judicial officers must not use generative artificial intelligence for judicial decision-making.
Supportive uses may organise case materials, summarise, aid research and proof-read. They are not a substitute for reading the evidence or for fact-finding.
Case-related use is limited to Court-approved tools.
This Practice Note was issued and commences on 14 May 2026.
This Practice Note applies to all Court users.
This Practice Note was issued replacing the Guidelines for Litigants: Responsible Use of Artificial Intelligence in Litigation of May 2024.
Court users can utilise AI in the preparation of Court documents if they comply with this Practice Note.
Content produced using AI must be verified with meaningful human control.
One AI tool cannot be used to confirm the content generated by another AI tool.
Particular caution must be exercised if GenAI tools are used to assist in the preparation of affidavit materials, witness statements or other documents which form the evidence or potential evidence of a witness.
Court users should not enter any information which is confidential or sensitive (including information subject to a non-publication or suppression order) into a public AI tool.
Use of AI must not indirectly mislead the Court or others about the nature of work or content; AI-assisted legal tasks remain subject to practitioners' duties including candour.
Paras 3-4.
The County Court of Victoria publishes equivalent litigant guidelines (per Law Society NSW protocol table, 14 May 2026).
Victorian Legal Services Board + Commissioner (with Uniform Law regulators)
06 December 2024 · Effective 2024-12-06
Dated 6 December 2024. Joint statement of the Law Society of New South Wales, the Legal Practice Board of Western Australia and the Victorian Legal Services Board and Commissioner. It applies existing Legal Profession Uniform Law duties; it does not create a new rule. Lawyers cannot safely enter confidential, sensitive or privileged client information into public AI chatbots or copilots such as ChatGPT, or any other public tool. Commercial tools used with client information require a careful review of the contract so the information will be kept secure (ASCR r 9.1; Barristers' Rules r 114). AI chatbots, copilots and other tools based on large language models cannot reason, understand or advise. The lawyer must exercise their own forensic judgment and must not treat the output as a substitute for their own assessment (ASCR r 4.1.4; Barristers' Rules rr 3(b), 4(e) and 42). No tool based on current large language models can be free of hallucinations. A lawyer who uses AI to prepare documents must be qualified to verify them and must actually ensure they are accurate and not likely to mislead the client, the court or another party (ASCR rr 4.1.2 and 4.1.3; Barristers' Rules rr 4(c)-(d), 8(a) and 35). Time and work items billed must accurately represent work done by the practice. AI use should not increase the client's costs above traditional methods, for example by extra time spent verifying or correcting output (Uniform Law ss 172-173; ASCR r 12.2). The statement also says lawyers using AI should consider risk-based policies on which tools may be used, by whom, for what, and with what information, including supervision of junior and support staff and verification before documents are settled; should limit use to lower-risk tasks that are easier to verify, with the examples of drafting a polite email or suggesting how to structure an argument; and should not use it for higher-risk tasks given as examples of translating advice into another language, analysing an unfamiliar legal concept, or executive decision-making. They should record and, where necessary or appropriate, disclose to the client, the court and fellow practitioners when and how AI was used, and how that use is reflected in costs if the client asks. The regulators say they will review the guidance as use evolves.
Confidentiality (ASCR r 9.1; BR r 114): do not enter confidential client information into public chatbots.
Independent advice (ASCR r 4.1.4).
Competence and honesty (ASCR rr 4.1.2-4.1.3).
Fair costs (Uniform Law ss 172-173; ASCR r 12.2): bill actual work done.
Limit AI to lower-risk, easier-to-verify tasks; be transparent with clients.
Victorian Legal Services Board and Commissioner publishes “New research examines AI use in the Victorian legal profession | VLSB+C”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board and Commissioner publishes “Generative AI Use in the Legal Profession – Research Brief”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board and Commissioner publishes “Tapping the potential of AI in legal practice: daiquiris, anyone? | VLSB+C”. New research finds that while generative AI is gaining ground in Victorian legal practice, its benefits remain narrowly conceived and unevenly distributed. Checked 2026-10-10. Read the linked page before relying on it.
New research finds that while generative AI is gaining ground in Victorian legal practice, its benefits remain narrowly conceived and unevenly distributed.
Risk Outlook 2026 - Using artificial intelligenceRegulator reportRegulatorPrimary sourceConfidence: highUnverified date"Last updated on 29 April 2026" shown on page.Secondary. Scope not stated in the stored source.
Victorian Legal Services Board + Commissioner
2026-04-29
VLSB+C's 2026 risk outlook names AI as a professional competency risk, encouraging responsible use while warning that unverified GenAI output breaches professional obligations and may lead to disciplinary action.
Almost 40% of Victorian lawyers use GenAI tools, per the Legal Services Research Centre
Lawyers must personally verify GenAI outputs, including reading cited sources, and not generate outputs they cannot verify
Principals should set clear GenAI policies, confirm tool security and confidentiality, and train staff
Failures to verify AI material are taken extremely seriously and may lead to regulatory or disciplinary action
Lawyers should check whether clients used GenAI to produce documents for their matter
Office of the Victorian Information Commissioner publishes “Artificial Intelligence and Privacy – Issues and Challenges”. Checked 2026-10-10. Read the linked page before relying on it.
Risk Outlook 2024 - Artificial intelligence: improper useRegulator reportRegulatorPrimary sourceConfidence: mediumUnverified datePage shows dates 6 June 2024 and a "last updated" of 27 March 2026; publication date not confirmed.Secondary. Guides lawyers. Not a court practice note.
Victorian Legal Services Board + Commissioner
Not stated
VLSB+C's 2024 risk outlook identifying improper use of AI as a key risk and listing the regulator's expectations of lawyers who use AI.
Lawyers are responsible for the accuracy of output, not the AI program
Never use AI to create outputs you cannot personally validate (e.g. translations into a language you do not speak)
Never enter confidential, sensitive or privileged client data into generic LLMs
Supervise junior lawyers and support staff using AI and verify their output
Office of the Victorian Information Commissioner publishes “Artificial Intelligence – Understanding Privacy Obligations”. Checked 2026-10-10. Read the linked page before relying on it.
Generative AI and LawyersRegulator statementRegulatorPrimary sourceConfidence: highUnverified dateHistoric (2023) statement; later superseded in practice by the Dec 2024 joint Statement on AI in legal practice.Secondary. Guides lawyers. Not a court practice note.
Victorian Legal Services Board + Commissioner
2023-09-28
Early VLSB+C commentary (28 September 2023) on ChatGPT and other generative AI platforms: how they work, accuracy limits and the ethical rules that apply when lawyers use them.
ChatGPT is a generalist tool, not trained to give accurate legal answers, and is optimised for fluency over accuracy
Lawyers remain responsible for the accuracy of any text they generate
Recommends lawyers gain a basic understanding of generative AI, its uses and risks
Office of the Victorian Information Commissioner publishes “Public consultation on Artificial Intelligence privacy guidance”. Checked 2026-10-10. Read the linked page before relying on it.
Office of the Victorian Information Commissioner publishes “Use of enterprise Generative AI tools in the Victorian public sector”. Checked 2026-10-10. Read the linked page before relying on it.
Law Institute of Victoria publishes “Artificial Intelligence Hub - Law Institute of Victoria”. Share your feedback to help us enhance our support on AI topics. Checked 2026-10-10. Read the linked page before relying on it.
Share your feedback to help us enhance our support on AI topics.
Approved by the CEO of the Law Institute of Victoria on 13 August 2025, after consultation with the VLSB+C and the LPLC. The guideline says these are guidelines only and do not have the force of law. They apply to practitioners and support staff in a law practice, and to other organisations that deliver legal services. Practitioners who use AI should: not enter confidential or legally privileged information into an open-source or commercial AI system without client consent; check all outputs for accuracy and suitability; check that cases exist and are relevant; tell the client before AI is used; be able to tell the opponent and the court, if asked, that AI was used; read the privacy policy and terms before first use and review them regularly; charge fairly; and retain what is needed for conflict checks after the retainer ends. The practitioner remains accountable for documents tendered and cannot delegate that to an AI system, staff or the client. The guideline says practitioners should be cautious until they have received appropriate training. Not every use must be disclosed if the text has been thoroughly checked, but the practitioner must know when a document was drafted with AI and be able to tell the court its origin. Tell the client before AI is used in the matter. Do not input confidential or privileged information except into a closed system whose claim to be closed has been carefully examined, and only after weighing the risk. Seek the client's consent before using confidential information with an AI model. An information barrier can be undermined if a closed system retains one client's information and later uses it for another. Principals are responsible for reasonable supervision, including of support staff who can access AI (ASCR r 37.1). A time-based bill cannot charge two hours for a document generated with AI in 20 minutes, including time spent instructing the AI and reviewing the result. Costs related to AI use are to appear in the costs disclosure and estimate.
Approved 13 August 2025. Guidelines only; they do not have the force of law.
Do not enter confidential or privileged information into an open or commercial AI system without client consent.
Tell the client before AI is used. Be able to tell the opponent and the court, if asked, whether AI was used.
Principals must supervise staff who can access AI.
Do not bill manual time for work generative AI shortened. The guideline's example is two hours billed for 20 minutes of work.
Same Law Institute of Victoria guidance note as the catalogue row for this PDF without the tracking parameter. Approved by the CEO on 13 August 2025, after consultation with the VLSB+C and the LPLC. The guideline says these are guidelines only and do not have the force of law. The 6 December 2024 date in the document is a footnote citation of the Uniform Law statement, not the approval date of this note.
Law Institute of Victoria publishes “Artificial Intelligence Hub - Law Institute of Victoria”. Share your feedback to help us enhance our support on AI topics. Checked 2026-10-10. Read the linked page before relying on it.
Share your feedback to help us enhance our support on AI topics.
Law Institute of Victoria publishes “Artificial Intelligence and Lawyers' Ethical Duties”. Checked 2026-10-10. Read the linked page before relying on it.
Artificial intelligence & large language models - Victorian BarProfessional guidanceSecondary sourceUnverified dateThe 10 October 2026 check marked this find as low confidence. Read the page before treating it as the body’s position.Secondary. Scope not stated in the stored source.
Victorian Bar
Not stated
Victorian Bar publishes “Artificial intelligence & large language models - Victorian Bar”. Checked 2026-10-10. Read the linked page before relying on it.
LIJ report on the Victorian Justice Leaders AI Forum convened by the LIV with the Victorian Attorney-General, and the published summary of enablers and actions for AI adoption in the justice sector.
More than 130 leaders from courts, tribunals, regulators and the profession attended
Focus on access to justice, productivity and integrity of the legal system
Summary of priority actions available on the LIV AI Hub; AI is a pillar of the LIV Call to Parties
Victoria Legal Aid publishes “Upgrading our online chat to provide a better service to help seekers”. The questions our digital assistant asks you are the same for every chat. The tool does not use artificial intelligence, think, or learn on its own. Checked 2026-10-10. Read the linked page before relying on it.
The questions our digital assistant asks you are the same for every chat. The tool does not use artificial intelligence, think, or learn on its own.
Victoria Legal Aid publishes “Making community legal information | Victoria Legal Aid”. Artificial intelligence (AI) can support community legal information work. We find it useful for some tasks, while recognising that it also has risks and limitations. Checked 2026-10-10. Read the linked page before relying on it.
Artificial intelligence (AI) can support community legal information work. We find it useful for some tasks, while recognising that it also has risks and limitations.
Victoria Legal Aid publishes “Protecting privacy and data – obligations for panel practitioners”. The way legal services are delivered is rapidly evolving, shaped by digital communication, remote work, and the growing use of artificial intelligence. Checked 2026-10-10. Read the linked page before relying on it.
The way legal services are delivered is rapidly evolving, shaped by digital communication, remote work, and the growing use of artificial intelligence.
VCAT expert-evidence practice note (effective 1 February 2026) that requires experts to be made aware of PNVCAT11 on Gen AI and to include a Gen AI disclosure in expert reports where relevant.
Party retaining an expert must make the expert aware of PNVCAT 11 (Use of Generative AI) where relevant
Expert reports must include, where relevant, a disclosure about Gen AI use as required by PNVCAT 11
VLRC final report, tabled on 3 February 2026, with 30 recommendations. It proposes a principles-based approach, eight principles for safe AI use, and guidelines for court users, judicial officers and staff, plus governance and training. The recommendations are not binding until a court or tribunal adopts them. Earlier batches hold the consultation paper, not this report.
30 recommendations to guide safe use of AI in Victorian courts and tribunals
Recommends a principles-based regulatory approach and eight principles to maintain public trust
Guidelines recommended for court users (including lawyers), judicial officers and court/tribunal staff, with an example of updated Supreme Court guidelines
Calls for formal, ongoing training and education on AI guidelines
This is the catalogue for Victoria. An instrument appears under “Rules that apply” only when you name the court or tribunal, or say that no court is involved.
Issued 5 June 2026. It supersedes the County Court’s 3 July 2024 litigant guidelines. It does not require a disclosure statement. Court users must be prepared to identify portions produced with AI and explain how the output was verified. The note uses its own paragraph numbering. It is not a copy of the Supreme Court note’s paragraph numbers.
Court users must be prepared to identify the specific portions of Court documents which were produced using AI and be able to explain how the output was verified.
The guideline PDF is dated 14 May 2026. · Effective 2026-05-14
These are guidelines for judges, associate judges and judicial registrars of the Supreme Court of Victoria, including reserves. They are not a practice note binding the parties. The PDF is dated 14 May 2026. They are informed by eight principles in the Victorian Law Reform Commission report of October 2025. Judicial officers must not use generative artificial intelligence for judicial decision-making. They exercise their own judgment and give reasons. The guidelines do not prevent supportive uses by officers or staff. Permissible supportive uses are organising and locating case materials, summaries and chronologies, an aid to legal research, and proof-reading. Those uses are not a substitute for reading or listening to the evidence and submissions, or for fact-finding. For case-related matters, only Court-approved tools may be used, with confidentiality controls. Officers should direct staff on use and disclosure. The guidelines will be reviewed. This is guidance for judicial officers.
Guidance for judicial officers, not a practice note binding the parties.
Judicial officers must not use generative artificial intelligence for judicial decision-making.
Supportive uses may organise case materials, summarise, aid research and proof-read. They are not a substitute for reading the evidence or for fact-finding.
Case-related use is limited to Court-approved tools.
VCAT expert-evidence practice note (effective 1 February 2026) that requires experts to be made aware of PNVCAT11 on Gen AI and to include a Gen AI disclosure in expert reports where relevant.
Party retaining an expert must make the expert aware of PNVCAT 11 (Use of Generative AI) where relevant
Expert reports must include, where relevant, a disclosure about Gen AI use as required by PNVCAT 11
VLRC final report, tabled on 3 February 2026, with 30 recommendations. It proposes a principles-based approach, eight principles for safe AI use, and guidelines for court users, judicial officers and staff, plus governance and training. The recommendations are not binding until a court or tribunal adopts them. Earlier batches hold the consultation paper, not this report.
30 recommendations to guide safe use of AI in Victorian courts and tribunals
Recommends a principles-based regulatory approach and eight principles to maintain public trust
Guidelines recommended for court users (including lawyers), judicial officers and court/tribunal staff, with an example of updated Supreme Court guidelines
Calls for formal, ongoing training and education on AI guidelines
Law Institute of Victoria publishes “Artificial Intelligence Hub - Law Institute of Victoria”. Share your feedback to help us enhance our support on AI topics. Checked 2026-10-10. Read the linked page before relying on it.
Share your feedback to help us enhance our support on AI topics.
Approved by the CEO of the Law Institute of Victoria on 13 August 2025, after consultation with the VLSB+C and the LPLC. The guideline says these are guidelines only and do not have the force of law. They apply to practitioners and support staff in a law practice, and to other organisations that deliver legal services. Practitioners who use AI should: not enter confidential or legally privileged information into an open-source or commercial AI system without client consent; check all outputs for accuracy and suitability; check that cases exist and are relevant; tell the client before AI is used; be able to tell the opponent and the court, if asked, that AI was used; read the privacy policy and terms before first use and review them regularly; charge fairly; and retain what is needed for conflict checks after the retainer ends. The practitioner remains accountable for documents tendered and cannot delegate that to an AI system, staff or the client. The guideline says practitioners should be cautious until they have received appropriate training. Not every use must be disclosed if the text has been thoroughly checked, but the practitioner must know when a document was drafted with AI and be able to tell the court its origin. Tell the client before AI is used in the matter. Do not input confidential or privileged information except into a closed system whose claim to be closed has been carefully examined, and only after weighing the risk. Seek the client's consent before using confidential information with an AI model. An information barrier can be undermined if a closed system retains one client's information and later uses it for another. Principals are responsible for reasonable supervision, including of support staff who can access AI (ASCR r 37.1). A time-based bill cannot charge two hours for a document generated with AI in 20 minutes, including time spent instructing the AI and reviewing the result. Costs related to AI use are to appear in the costs disclosure and estimate.
Approved 13 August 2025. Guidelines only; they do not have the force of law.
Do not enter confidential or privileged information into an open or commercial AI system without client consent.
Tell the client before AI is used. Be able to tell the opponent and the court, if asked, whether AI was used.
Principals must supervise staff who can access AI.
Do not bill manual time for work generative AI shortened. The guideline's example is two hours billed for 20 minutes of work.
Same Law Institute of Victoria guidance note as the catalogue row for this PDF without the tracking parameter. Approved by the CEO on 13 August 2025, after consultation with the VLSB+C and the LPLC. The guideline says these are guidelines only and do not have the force of law. The 6 December 2024 date in the document is a footnote citation of the Uniform Law statement, not the approval date of this note.
Use of AI must not indirectly mislead the Court or others about the nature of work or content; AI-assisted legal tasks remain subject to practitioners' duties including candour.
Paras 3-4.
The County Court of Victoria publishes equivalent litigant guidelines (per Law Society NSW protocol table, 14 May 2026).
Law Institute of Victoria publishes “Artificial Intelligence Hub - Law Institute of Victoria”. Share your feedback to help us enhance our support on AI topics. Checked 2026-10-10. Read the linked page before relying on it.
Share your feedback to help us enhance our support on AI topics.
Law Institute of Victoria publishes “Artificial Intelligence and Lawyers' Ethical Duties”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board + Commissioner (with Uniform Law regulators)
06 December 2024 · Effective 2024-12-06
Dated 6 December 2024. Joint statement of the Law Society of New South Wales, the Legal Practice Board of Western Australia and the Victorian Legal Services Board and Commissioner. It applies existing Legal Profession Uniform Law duties; it does not create a new rule. Lawyers cannot safely enter confidential, sensitive or privileged client information into public AI chatbots or copilots such as ChatGPT, or any other public tool. Commercial tools used with client information require a careful review of the contract so the information will be kept secure (ASCR r 9.1; Barristers' Rules r 114). AI chatbots, copilots and other tools based on large language models cannot reason, understand or advise. The lawyer must exercise their own forensic judgment and must not treat the output as a substitute for their own assessment (ASCR r 4.1.4; Barristers' Rules rr 3(b), 4(e) and 42). No tool based on current large language models can be free of hallucinations. A lawyer who uses AI to prepare documents must be qualified to verify them and must actually ensure they are accurate and not likely to mislead the client, the court or another party (ASCR rr 4.1.2 and 4.1.3; Barristers' Rules rr 4(c)-(d), 8(a) and 35). Time and work items billed must accurately represent work done by the practice. AI use should not increase the client's costs above traditional methods, for example by extra time spent verifying or correcting output (Uniform Law ss 172-173; ASCR r 12.2). The statement also says lawyers using AI should consider risk-based policies on which tools may be used, by whom, for what, and with what information, including supervision of junior and support staff and verification before documents are settled; should limit use to lower-risk tasks that are easier to verify, with the examples of drafting a polite email or suggesting how to structure an argument; and should not use it for higher-risk tasks given as examples of translating advice into another language, analysing an unfamiliar legal concept, or executive decision-making. They should record and, where necessary or appropriate, disclose to the client, the court and fellow practitioners when and how AI was used, and how that use is reflected in costs if the client asks. The regulators say they will review the guidance as use evolves.
Confidentiality (ASCR r 9.1; BR r 114): do not enter confidential client information into public chatbots.
Independent advice (ASCR r 4.1.4).
Competence and honesty (ASCR rr 4.1.2-4.1.3).
Fair costs (Uniform Law ss 172-173; ASCR r 12.2): bill actual work done.
Limit AI to lower-risk, easier-to-verify tasks; be transparent with clients.
Artificial intelligence & large language models - Victorian BarProfessional guidanceSecondary sourceUnverified dateThe 10 October 2026 check marked this find as low confidence. Read the page before treating it as the body’s position.Secondary. Scope not stated in the stored source.
Victorian Bar
Not stated
Victorian Bar publishes “Artificial intelligence & large language models - Victorian Bar”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board and Commissioner publishes “New research examines AI use in the Victorian legal profession | VLSB+C”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board and Commissioner publishes “Generative AI Use in the Legal Profession – Research Brief”. Checked 2026-10-10. Read the linked page before relying on it.
Victorian Legal Services Board and Commissioner publishes “Tapping the potential of AI in legal practice: daiquiris, anyone? | VLSB+C”. New research finds that while generative AI is gaining ground in Victorian legal practice, its benefits remain narrowly conceived and unevenly distributed. Checked 2026-10-10. Read the linked page before relying on it.
New research finds that while generative AI is gaining ground in Victorian legal practice, its benefits remain narrowly conceived and unevenly distributed.
Victoria Legal Aid publishes “Upgrading our online chat to provide a better service to help seekers”. The questions our digital assistant asks you are the same for every chat. The tool does not use artificial intelligence, think, or learn on its own. Checked 2026-10-10. Read the linked page before relying on it.
The questions our digital assistant asks you are the same for every chat. The tool does not use artificial intelligence, think, or learn on its own.
Victoria Legal Aid publishes “Making community legal information | Victoria Legal Aid”. Artificial intelligence (AI) can support community legal information work. We find it useful for some tasks, while recognising that it also has risks and limitations. Checked 2026-10-10. Read the linked page before relying on it.
Artificial intelligence (AI) can support community legal information work. We find it useful for some tasks, while recognising that it also has risks and limitations.
Victoria Legal Aid publishes “Protecting privacy and data – obligations for panel practitioners”. The way legal services are delivered is rapidly evolving, shaped by digital communication, remote work, and the growing use of artificial intelligence. Checked 2026-10-10. Read the linked page before relying on it.
The way legal services are delivered is rapidly evolving, shaped by digital communication, remote work, and the growing use of artificial intelligence.
Risk Outlook 2026 - Using artificial intelligenceRegulator reportRegulatorPrimary sourceConfidence: highUnverified date"Last updated on 29 April 2026" shown on page.Secondary. Scope not stated in the stored source.
Victorian Legal Services Board + Commissioner
2026-04-29
VLSB+C's 2026 risk outlook names AI as a professional competency risk, encouraging responsible use while warning that unverified GenAI output breaches professional obligations and may lead to disciplinary action.
Almost 40% of Victorian lawyers use GenAI tools, per the Legal Services Research Centre
Lawyers must personally verify GenAI outputs, including reading cited sources, and not generate outputs they cannot verify
Principals should set clear GenAI policies, confirm tool security and confidentiality, and train staff
Failures to verify AI material are taken extremely seriously and may lead to regulatory or disciplinary action
Lawyers should check whether clients used GenAI to produce documents for their matter
Office of the Victorian Information Commissioner publishes “Artificial Intelligence and Privacy – Issues and Challenges”. Checked 2026-10-10. Read the linked page before relying on it.
Risk Outlook 2024 - Artificial intelligence: improper useRegulator reportRegulatorPrimary sourceConfidence: mediumUnverified datePage shows dates 6 June 2024 and a "last updated" of 27 March 2026; publication date not confirmed.Secondary. Guides lawyers. Not a court practice note.
Victorian Legal Services Board + Commissioner
Not stated
VLSB+C's 2024 risk outlook identifying improper use of AI as a key risk and listing the regulator's expectations of lawyers who use AI.
Lawyers are responsible for the accuracy of output, not the AI program
Never use AI to create outputs you cannot personally validate (e.g. translations into a language you do not speak)
Never enter confidential, sensitive or privileged client data into generic LLMs
Supervise junior lawyers and support staff using AI and verify their output
Office of the Victorian Information Commissioner publishes “Artificial Intelligence – Understanding Privacy Obligations”. Checked 2026-10-10. Read the linked page before relying on it.
Generative AI and LawyersRegulator statementRegulatorPrimary sourceConfidence: highUnverified dateHistoric (2023) statement; later superseded in practice by the Dec 2024 joint Statement on AI in legal practice.Secondary. Guides lawyers. Not a court practice note.
Victorian Legal Services Board + Commissioner
2023-09-28
Early VLSB+C commentary (28 September 2023) on ChatGPT and other generative AI platforms: how they work, accuracy limits and the ethical rules that apply when lawyers use them.
ChatGPT is a generalist tool, not trained to give accurate legal answers, and is optimised for fluency over accuracy
Lawyers remain responsible for the accuracy of any text they generate
Recommends lawyers gain a basic understanding of generative AI, its uses and risks
Office of the Victorian Information Commissioner publishes “Public consultation on Artificial Intelligence privacy guidance”. Checked 2026-10-10. Read the linked page before relying on it.
Office of the Victorian Information Commissioner publishes “Use of enterprise Generative AI tools in the Victorian public sector”. Checked 2026-10-10. Read the linked page before relying on it.
LIJ report on the Victorian Justice Leaders AI Forum convened by the LIV with the Victorian Attorney-General, and the published summary of enablers and actions for AI adoption in the justice sector.
More than 130 leaders from courts, tribunals, regulators and the profession attended
Focus on access to justice, productivity and integrity of the legal system
Summary of priority actions available on the LIV AI Hub; AI is a pillar of the LIV Call to Parties
This Practice Note was issued and commences on 14 May 2026.
This Practice Note applies to all Court users.
This Practice Note was issued replacing the Guidelines for Litigants: Responsible Use of Artificial Intelligence in Litigation of May 2024.
Court users can utilise AI in the preparation of Court documents if they comply with this Practice Note.
Content produced using AI must be verified with meaningful human control.
One AI tool cannot be used to confirm the content generated by another AI tool.
Particular caution must be exercised if GenAI tools are used to assist in the preparation of affidavit materials, witness statements or other documents which form the evidence or potential evidence of a witness.
Court users should not enter any information which is confidential or sensitive (including information subject to a non-publication or suppression order) into a public AI tool.
The privacy and freedom-of-information regime for Victoria, and what the 10 October 2026 check recorded about AI. Confirm the statute, and any commencement proclamation, before reliance.
Privacy and Data Protection Act 2014 (IPPs; Victorian Protective Data Security Standards); Health Records Act 2001. FOI: Freedom of Information Act 1982 (Vic). OVIC guidance covers AI privacy obligations and enterprise GenAI tools in the Victorian public sector. VLA tells panel practitioners how to protect privacy and data. Checked 2026-10-10.
Privacy and Data Protection Act 2014 (IPPs; Victorian Protective Data Security Standards); Health Records Act 2001
FOI: Freedom of Information Act 1982 (Vic)
OVIC: AI – understanding privacy obligations — https://ovic.vic.gov.au/privacy/resources-for-organisations/artificial-intelligence-understanding-privacy-obligations/
OVIC: Enterprise GenAI tools in the Victorian public sector — https://ovic.vic.gov.au/privacy/resources-for-organisations/use-of-enterprise-generative-ai-tools-in-the-victorian-public-sector/
VLA: Privacy and data obligations for panel practitioners — https://www.legalaid.vic.gov.au/protecting-privacy-and-data-security-obligations-panel-practitioners
Statute names from official sources; how they apply to AI is a summary to verify.
Court users are responsible for the content whether or not a tool was used. Lawyers must exercise oversight and verify accuracy. The note warns of referral to the Victorian Legal Services Board and Commissioner, and of personal costs orders, if unverified output is relied on.
Public tools and confidential facts do not mix
The note says court users should not enter confidential or sensitive information, including material under a suppression or non-publication order, into a public AI tool. Closed tools still have to be checked against their terms.
Evidence and experts need extra care
Affidavit material must reflect the person’s own knowledge and words. Expert reports must comply with the applicable code or practice note. The note requires particular caution in both cases.
Jurisdiction
SC Gen 25 is a Supreme Court note. The County Court has published its own litigant guidelines. VCAT is a different forum. Confirm that tribunal’s current direction rather than assuming SC Gen 25 applies.
What to open first
If the document is going to the Supreme Court of Victoria, SC Gen 25 is the first thing to open. Then verify the contents in a source that is not another chatbot.
Review checklist
Open SC Gen 25 before a Supreme Court document is filed.
Fact-check, proofread and adapt any generated content. Verify authorities in AustLII, an official legislation site or the Law Library of Victoria, as the note describes.
Do not ask an AI tool to confirm another AI tool’s citations.
Keep confidential and suppressed material out of public tools.
Be able to identify which parts of a court document were produced with AI and how they were verified.
Jurisdiction-specific considerations
Verification by a second model is expressly not enough.
A closed tool is not automatically safe. The note says you still look at the contractual and privacy settings.
The County Court and VCAT are not the Supreme Court. Check their own material.
Limits on AI assistance
Do not file, send or rely on generated text until the responsible lawyer has checked it against the instructions and the current source.
Do not put suppressed, privileged, subpoenaed or safety-sensitive information into a public AI tool.
Check the original testamentary documents Confirm the correct applicant and grant type Check the inventory and current filing requirements before completing the application
