Supreme Court of Victoria issues AI guidelines for court users and judicial officers
The Supreme Court of Victoria yesterday released a new Practice Note (SC Gen 25) on the use of artificial intelligence (AI) by court users, together with separate Judicial Guidelines for judicial officers. The court was one of the first jurisdictions in Australia to develop AI guidelines back in May 2024.
The freshly released Practice Note and Judicial Guidelines provide a much-needed update – further clarifying the use of AI and the effect of this emerging technology on the judiciary following the extensive analysis of the use of AI in Victoria’s court and tribunals undertaken recently by the Victorian Law Reform Commission in its Report.
The Practice Note follows the recent release of the Federal Court of Australia’s own ‘Use of Generative Artificial Intelligence Practice Note’ (see our previous article, ‘Federal Court releases Use of Generative AI Practice Note: key guidance for using AI in proceedings’ for more on this). While the Federal Court’s guidance noted the Commission’s Report on AI; the Practice Note and Judicial Guidelines go further to more strongly adopt the principles and recommendations formulated by the Report.
This updated position represents a significant step forward in providing courts, tribunals and people interacting with the justice system coherent guidance on the appropriate use of AI.
Strong alignment with the Report
The Victorian Law Reform Commission’s ‘Artificial Intelligence in Victoria’s Courts and Tribunals Report’ (tabled in February 2026) was the first inquiry by an Australian law reform body into the use of AI in courts and tribunals. We have previously discussed the significance of this extensive Report, and the breadth of its recommendation for the judiciary, in our Smarter Law podcast series, ‘AI and legal professional privilege’.
The Report consulted with Victoria’s courts, VCAT, the legal profession and community organisations and found that fewer than a third of Victorian lawyers surveyed had read the existing 2024 guidelines and some 80 per cent of lawyers agreed that insufficient guidance on responsible AI use posed a risk to legal practice.
The Report recommended reforms to the Supreme Court’s guidelines, including clearer definitions, broader application, specific direction on submissions, affidavits and expert reports, and a formal principles framework. It also recommended separate publicly available guidelines for judicial officers.
The Practice Note and the Judicial Guidelines explicitly adopt many of these recommendations in what is a strong endorsement of the Commission’s Report.
Definitions adopted verbatim
The 2024 guidelines described AI more broadly as ‘a term describing a range of technologies and techniques used to computationally generate outputs that typically require human intelligence to produce.’
In comparison, the Practice Note now adopts the more detailed definition of AI in the Commission’s Report and very significantly defines Generative AI, public AI and closed AI as distinct subcategories. The Report strongly recommended that the 2024 guideline be updated to reflect these important definitions.
The distinction between public and closed AI was absent from the 2024 guidelines. It is critical for practitioners to have a clear understanding of this distinction to manage confidentiality and privilege risks with respect to the use of AI.
Scope expanded
The 2024 guidelines were directed to litigants and did not cover witnesses or experts. The Practice Note defines ‘Court user’ to include lawyers, litigants (whether legally represented or self-represented), and lay and expert witnesses, in line with the recommendations of the Report.
Principles framework
The 2024 guidelines contained some informal principles about understanding AI limitations and disclosing use. The Practice Note now incorporates all eight of the principles proposed by the Report to guide the safe use of AI, namely:
- access to justice;
- impartiality and fairness
- accountability and independence;
- privacy and data security;
- transparency and open justice;
- contestability and procedural fairness;
- efficiency and effectiveness; and
- human oversight and monitoring.
Verification requirements for the accuracy of submissions
The Report recommended the 2024 guidelines be updated with respect to material submitted to the court. To ensure accuracy, court users should:
- fact-check and proofread;
- edit and adapt;
- verify references;
- use sources known to be accurate; and
- not use GenAI to verify GenAI outputs.
The Practice Note adopts this wording almost verbatim. The 2024 guidelines encouraged checking, while the Practice Note states these steps in directive terms.
The Practice Note also specifies that ‘[o]ne AI tool cannot be used to confirm the content generated by another AI tool’, rather, courts users must maintain vigilance, with the Practice Note reinforcing that ‘[c]ontent produced using AI must be verified with meaningful human control’.
Affidavits and witness statements
The Report recommended that a person making a witness statement or similar statement must verify its accuracy and that it represents their view.
The Practice Note implements this by requiring that documents be ‘sworn/affirmed or finalised in a form that reflects that person’s own knowledge and words.’ The focus is on accuracy verification rather than mandatory disclosure of AI use in this regard. This is consistent with the Commission’s comments in the Report that ‘disclosure is less meaningful than checking accuracy.’
Expert reports
The Practice Note requires compliance with ‘the relevant Code of Conduct or Practice Note’ for expert reports, cross-referencing Practice Note SC CR 3, which requires experts to disclose AI use and identify any possible biases that may affect the content generated by the AI tool used. The Commission’s Report called for exactly this cross-reference.
Judicial Guidelines
The Commission’s Report called for public guidelines for all judicial officers that prohibit AI for judicial decision-making, clarify that supportive uses are permitted, and encourage staff to discuss AI use with their supervising judicial officer.
The new Judicial Guidelines explicitly implement these recommendations.
Dealing with confidential and privileged information
The developing commentary on AI’s impact on legal professional privilege and confidentiality increasingly notes the critical issue of whether an AI tool is ‘public/open’ or ‘closed’, and how that classification affects legal professional privilege.
Notably, the Practice Note distinguishes between ‘public’ and ‘closed’ AI tools, something the 2024 guidelines did not do.
For public AI tools, defined as AI tools ‘openly accessible to the public’, court users must not enter any confidential or sensitive information, including information subject to non-publication or suppression orders. Lawyers specifically must not enter privileged client information into public AI tools. Adherence to a lawyer’s professional obligations to maintain client confidentiality is enshrined in the Practice Note.
For closed AI tools, defined as including ‘controls to reduce risks’, users must review the contractual terms and privacy settings to satisfy themselves that information will remain secure and not be used to train the AI program.
This directly addresses the warning in the Commission’s Report on the risk of inadvertent waiver of legal professional privilege when information is uploaded to public AI models.
The Federal Court’s recent AI guidance takes a similar approach, warning that privileged, confidential or private information must not be used in a way inconsistent with the obligations applying to that information. Both instruments reflect the same underlying concern. The ease of inputting case materials into AI tools creates a real and immediate privilege risk that practitioners have to actively manage.
Seen in the broader context, these developments reinforce that AI’s impact on legal professional privilege is not a technicality or hypothetical. It is a current and material risk, requiring deliberate proactive decision making and ongoing vigilance.
Significance and next steps
The issuance of the Practice Note and the Judicial Guidelines represents the most significant recent development in AI governance for Victorian courts.
The instruments directly implement many of the Victorian Law Reform Commission’s key recommendations within three months of the Report being tabled in Parliament. The elevation from informal guidelines to a formal Practice Note means non-compliance now carries consequences. The court may take a failure to comply into account when exercising case management and costs powers, and lawyers who rely on unverified AI outputs risk referral to the Victorian Legal Services Board and Commissioner.
Some reforms remain outstanding, though these are largely as a result of them applying to all courts and tribunals in Victoria more broadly. For example, the Expert Witness Code of Conduct for civil trials has not yet been formally amended. Victoria’s other courts and VCAT have not yet adopted the Supreme Court’s Practice Note. Further, the governance structures recommended by the Commission, including a cross-jurisdictional technology and innovation committee, likely fall within Court Services Victoria’s responsibility to implement.
What remains unclear is how the judiciary will interpret and enforce these new instruments in practice.
The questions that remain are both numerous, and practical. For example, how much verification is enough? Will the court draw a distinction between a practitioner who used AI carelessly and one who used it competently but without disclosing such use? Further guidance will emerge only when the court directly considers these issues.
Considering the speed in which regulators and courts domestically are providing guidance on the growing use of AI in legal practice, and the international decisions that have started emerging in this space, it is only a matter of time before we see further developments in Australia. Watch this space.
This article was written with assistance from Marcus Jones, Law Graduate.
Contacts
