The rise of Artificial Intelligence in Litigation

With growing focus on improved practice efficiency, the vast quantities of analytical data, evidence and case law that practitioners are expected to review is objectively overwhelming, in its face the use of legal technology has become necessary to help ease the load and make sense of the sheer amount of data.

A recent report, Generative AI Use in the Legal Profession: Findings from the 2025 Victorian Lawyer Census, examines how Victorian practitioners are using AI in their legal practices and identifies the key challenges and risks associated with its use. The report found that 36.7% of respondents use AI tools in their legal practice, with more than half doing so weekly or daily. Adoption varied by years of experience, organisation type, and practice area.

AI software was most commonly used for background research on non-legal issues, case analysis and summaries, legal research, and client communication. It was used less often for document review, transcription, and contract drafting. Overall, practitioners primarily used AI for information gathering, drafting, and administrative tasks, rather than legal decision-making or preparing court documents.

Victorian Legal Services Board CEO and Commissioner Fiona McLeay observed that lawyers are currently using AI mainly to improve workflows and productivity, while its broader potential to enhance client services remains largely untapped. This reflects an ongoing debate about the appropriate role of AI in legal practice.

This year, Victorian Courts have revisited their AI guidelines. In April 2026, the Federal Court released its Use of Generative Artificial Intelligence Practice Note, and the Supreme Court issued their guidance in May. Most recently, the County Court issued their practice note on AI in June, together with separate guidelines for Judicial Officers. Such guidance follows the recommendations of the Victorian Law Reform Commission, issued in their report titled ‘Artificial Intelligence in Victoria’s Courts and Tribunals’ which was tabled in February 2026. The Practice Note’s offer clear guidance for litigants making clear that false or inaccurate material must not be presented to the Court and warns users to exercise caution when entering confidential or privileged information into AI tools.

The NSW State and District Courts have taken potentially the strongest stand on the use of AI whereby Generative AI must not be used to generate contents of affidavits, witness statements or character references and any such documents must contain a disclosure. Victoria guidelines however, as stated in the Supreme Court’s SCGEN 25 Practice Note, allow such use, but suggest the exercise of particular caution when dealing with the preparation of evidence.

The Victorian Practice Note recognises the potential of AI to enhance justice practices. It does allow for the use of AI while simultaneously detailing the risks court users should be aware of when using AI to prepare court documents and the steps to take in ensuring all information is complete, accurate and relevant to the case. Informed by recommendations from the Victorian Law Reform Commission, the rules expand the definition of court users explicitly to encompass lawyers, self-represented litigants and expert witnesses.

For legal practitioners and litigants, the guidelines mandate absolute personal accountabilities for all submitted material, strictly dictating that utilising AI does not absolve the user from ensuring the court documents remain complete, accurate and relevant. Filing a document that relies on unverified AI outputs may trigger severe consequences, including personal costs orders against practitioners and professional misconduct sanctions.

To prevent data leakage, users must differentiate between public and closed AI platforms, ensuring that sensitive case material remains entirely within controlled, secure environments. Additionally, expert witnesses must adhere to cross-referenced conduct codes requiring them to disclose any automated tool usage and identify potential structural biases in the generated data.

AI hallucinations and unreliable authorities

A key risk of using AI in legal practice is that it may produce information that is inaccurate, misleading or entirely fictitious. This can include fabricated case law, non-existent legislation and false legal citations. These errors are commonly described as “hallucinations”. Because large language models generate responses by predicting likely words rather than verifying legal authority, their outputs can appear persuasive even when they are wrong. For a profession grounded in accuracy, precedent and professional judgment, this presents a serious risk.

The issue has already arisen in a growing number of cases. In Dayal [2024], Mr Dayal tendered a list and summary of authorities that had been generated using AI and contained inaccurate citations and summaries. The Court acknowledged that technology is now an important part of efficient modern legal practice but cautioned that the opportunities offered by AI are accompanied by significant risks.

The consequences for Mr Dayal were significant. He was no longer entitled to practise as a principal lawyer, was no longer authorised to handle trust money or operate his own legal practice and was required to complete two years of supervised legal practice.

A similar issue arose in a Victorian murder trial involving a 14-year-old accused. The trial was halted after lawyers filed submissions that contained misleading AI-generated material. The incident underscored that the court’s ability to rely on the accuracy of submissions made by counsel is fundamental to the administration of justice.

The Court was told that the King’s Counsel and his junior solicitor had not properly checked the submissions before filing them. Those submissions included fictional case citations and inaccurate quotations from a parliamentary speech. The Supreme Court of Victoria guidelines make clear that AI must not be used in a way that misleads another participant in the litigation process about the work undertaken or the content produced. Of particular concern was counsel’s explanation that, after checking some initial citations, they wrongly assumed the remaining material would also be accurate.

 In his judgment, Justice Elliott reminded parties that, under the Supreme Court of Victoria guidelines, AI-generated material must be independently and thoroughly verified before it is used. The cases highlight the risk that AI can generate inaccurate or misleading information, and that reliance on unverified outputs can lead to serious professional and legal consequences.

Confidentiality and Privilege Risks

 Legal professional privilege protects confidential communications between lawyers and clients. Public AI tools may place that protection at risk if privileged or sensitive information is uploaded into them. Once entered, that material may be stored, used to train the system, disclosed to third parties or exposed through security vulnerabilities.

Generative AI may also create issues under non-disclosure agreements or client terms that restrict the use of AI technologies. Privileged or sensitive material should therefore not be entered into open AI platforms unless the platform is secure, confidential and approved for that purpose.

Although there is no direct Australian authority on this issue, United States v Heppner illustrates the approach Australian courts may take, as it reflects orthodox privilege principles. In that case, the defendant used a public AI platform to generate material about legal issues and defence strategy after becoming the target of a criminal investigation. When the government sought access to that material, the defendant claimed attorney-client privilege, but the claim was rejected.

The court held that users do not have a substantial privacy interest in conversations voluntarily disclosed to a publicly accessible AI platform and retained in the ordinary course of business. Without confidentiality, a claim for privilege is unlikely to arise. The case reinforces the need for clients and legal practitioners to treat AI platforms with caution and to preserve confidentiality as a central professional obligation.

As AI continues to develop faster than regulation can respond, legal practitioners and organisations should not wait for further rules before acting. Clear AI governance including staff training and updating employment policies to impose positive obligations on staff to declare use of AI in their work and verification processes, will be essential to ensure that AI is used safely, ethically and consistently with professional obligations. Practitioners should be cautious when using AI, ensure everything is verified and check the Court’s guidelines within the relevant jurisdiction.

Used carefully, AI has the potential to make legal services more efficient, accessible and responsive. However, AI and its value will depend on how deliberately the profession manages its risks and ensures that its benefits flow beyond internal productivity gains to clients, courts and the broader administration of justice.

At Sanicki Lawyers, we assist clients in navigating the legal and practical issues that arise in litigation, including the responsible use of emerging technologies such as artificial intelligence. Our experienced team can advise on obligations, risks and strategic considerations when AI is used in court-related processes.

Get in touch with us to discuss your rights and protections.

+61 3 9510 9888


hello@sanickilawyers.com.au