Tuesday 26 May 2026
Introduction
I begin by echoing the important acknowledgements of country and of National Sorry Day. I also pay my respects to the Traditional Owners of this land, the Wurundjeri People of the Kulin nation, their Elders, past and present, and to all First Nations people here this morning. As we meet to discuss innovation, which has serious implications for our environment, I am conscious that we do so on country which has been cared for, and has been home to innovation, for tens of thousands of years.
My thanks to the Law Institute of Victoria (LIV) for organising today’s forum, especially its CEO, Adam Awty and President, Tom Ballantyne, and to the Honourable Sonya Kilkenny MP, Attorney-General, for partnering with the LIV on this important initiative.
I am delighted to see such a broad section of the Victorian legal community represented here this morning – including representatives from across our Courts and Tribunals, regulators, private practice, public sector, government, corporate and in-house lawyers, and technology providers.
This morning, I want to touch on three themes: governance, investment and benefit. They are not novel but each of them is important. I want to briefly address their importance and explain how they are relevant to the Courts.
Before coming to them, it is important to highlight the context, for both the profession and the Courts, in which these themes arise. The rise and rise of AI should make us reflect on our role and function so that AI becomes part of, but does not overtake, our professional lives.
The key attributes of the legal profession on which clients depend are knowledge and judgment, both of which exist in an ethical setting. The fiduciary relationship between client and lawyer is an intrinsically human one and one built on trust. Thus, in order to receive legal advice, clients give the most confidential and sensitive information to their lawyers, confident that its privacy will be protected. The profession cannot exist nor justify its position unless it has these features in balance.
AI will improve access to information, and make it available in a very useful form, but AI is not a substitute for human judgment, and bringing AI into the circle of confidence carries clear risks including as to the location and access to the data.
To the extent that the data sources AI uses contain ethically informed decisions and judgments of the past, those attributes may be reflected in the output predicted by AI. But that process mimics – it does not equate with – ethical human judgment in the application of the law to individual cases.
Notwithstanding its tremendous but still incipient capacity, I do not think that AI will provide any sort of substitute for these cardinal features of the profession.
The Courts also depend on these features but, perhaps even more than the profession, they require the trust and confidence of the community they serve. Community trust and confidence in the judicial and the legal function is based on accountability, a trait founded on a conscious human choice.
The adoption of AI presents real challenges on this front.
The National Center for State Courts in the United States of America do an annual public opinion survey on attitudes to the State Courts. In the 2025 survey, a majority of people surveyed thought that AI would be more harmful than helpful to the Courts. 51 per cent agreed with the statement that ‘AI will hurt state courts by increasing the risk of mistakes that judges and staff can’t always catch and making it harder for people to trust court decisions.’ 31 per cent thought it would improve efficiency of Courts and 18 per cent did not know.1 Reporting of deep fakes, hallucinations and general suspicion of big tech may increase this level of mistrust. It is something that I expect would be reflected in our own State and we must demonstrate leadership on AI that accommodates these concerns.
I turn then to the themes, starting with governance.
Governance
By governance I mean the ethical and practical rules and frameworks within which AI is acquired and deployed.
It is tempting to see governance as a handbrake on innovation and a barrier to achieving the promised benefits of AI. But governance is crucial to progress for a number of reasons.
First, it would be imprudent to treat AI as just another technological development. Although AI does not share human characteristics, and remains anchored in algorithms that seek to predict the probability of various outcomes, it enables expression that mimics human interaction. It performs tasks that have hitherto been the domain of humans with incredible speed and accuracy.
Second, most end users will at best have only a rudimentary understanding about how the tools work and their strengths and vulnerabilities. Yet as lawyers and judges, each of us remain personally responsible for our work.
Third, there is an understandable temptation to see AI as a panacea for all the annoying parts of a job and it would be unwise to think that each individual user will bring an appropriately critical mind to each task. And the sycophantic approach of the LLMS is likely to seduce people to use AI at an increasing rate.
Fourth, the risks to the individual professional or judicial officer and to their firms, organisations or Courts of mistakes are enormous. If the community perceives that AI can provide a better but cheaper answer to lawyers or if judges use AI to make their decisions, their underlying rationale will fail. The value of human decision-making will be lost.
What are the Courts doing about it?
Courts Council, which comprises the heads of jurisdiction and two independent non-judicial members, have established an AI Committee, which is in the process of formulating a Court specific framework for the acquisition and deployment of AI. Our belief and expectation is that a court centred framework will best protect the unique features of the judicial system while driving innovation. In the meantime, we have a number of policies and protocols in place.
In 2024, the Supreme Court was one of the first Australian jurisdictions to develop AI guidelines.
On 14 May this year, the Court published two new documents – a Practice Note for Court Users2 and Guidelines for Judicial Officers.3 These documents supersede the earlier guidelines and have been developed to further clarify the use of AI following the release of the Victorian Law Reform Commission (VLRC)’s Report on AI in Victoria’s Courts and Tribunals.
The Court recognises the importance of good governance of emerging technology and of updating guidance and parameters as required. We hope that setting out the Court’s approach to the use of AI, and our requirements and expectations of Court users, will continue to increase confidence in the technology and its appropriate use.
Our Practice Note on the use of AI by Court users has been informed by the principles articulated by the VLRC in its report, including access to justice, impartiality, transparency and procedural fairness.
Court users can utilise AI in the preparation of Court documents if they comply with the Practice Note, which highlights the limitations and risks of AI, and what needs to be done to address them. Whether or not AI is used, Court users are responsible for the content of their Court documents, which must be verified with meaningful human control.
The Court’s Guidelines for Judicial Officers prohibit the use of Generative AI for judicial decision-making. However, Judicial Officers and Court staff can use AI in certain ways to support their work, including to organise and locate case materials, produce summaries and chronologies from case materials, as an aid to legal research and for proof-reading. These uses are not a substitute for reading and/or listening to evidence and submissions, or fact-finding when that is called for in judicial decision-making.
For case related matters, only Court-approved AI tools can be used. And Judicial Officers should provide the staff assisting them with clear directions regarding the appropriate use of AI and the level of disclosure the Judicial Officer requires.
Various Courts across State and Federal jurisdictions have issued guidance on the use of AI. For the most part, they are reasonably consistent in their approach – non-prescriptive, permissive, cautionary. As AI and its use cases evolve, the Supreme Court will continue to adapt its practice to ensure that we remain at the forefront of this developing area without sacrificing impartiality, privacy, accountability and fairness. Looking ahead, national consistency is desirable. In practice, I expect this will be an organic process.
Investment
The rate of investment in AI across the globe is truly staggering. Melbourne is a top-ten data centre market in the Asia-Pacific region,4 and adoption of AI is widespread as its capability continues to accelerate.
Investment will be in technology: the adoption of tools, some of which will be curated to the legal profession, is inevitable and happening at a fast rate. It is likely that some investment will not reap the promised benefits and there will be unavoidable losses. That always comes with change.
As important will be investment in people, through education and training. It is here that again, the organising principles which are intrinsic to the profession become important. The legal profession is not just a business and investment should not be considered based on the bottom line alone. There must be investment in instructing on the safe and ethical use of AI but as importantly, there must be a focus on developing analytical skills and in honing judgment in our junior practitioners. It will be in this domain, more than how to build a library of prompts, that the future of the profession depends.
To date, the Courts’ investment has been cautious and undertaken through pilot programs.
In March last year, an AI pilot program commenced in the Coroner’s Court to trial the potential of AI to increase the efficiency of reviewing and analysing case materials, as well as to reduce the risk of vicarious trauma by limiting instances of Court staff being unnecessarily exposed to distressing material.
In the Supreme Court, we are currently undertaking AI pilots to assess the potential of the technology to assist with different tasks in a range of contexts by a cross-section of Judicial Officers and Court staff.
Some general observations emerging from the experience so far include:
- Finding greater value through targeted tasks, as user skills develop, and when the product is deployed across material that is familiar to the user.
- Reducing the time to get across material and return to it, not by substituting summaries for reading, but by aiding the read in-process and reducing the time to navigate to known material.
- Avoiding using AI for tasks with a substantial warranting burden and instead targeting:
- tasks that do not require 100% accuracy; and
- tasks that produce an output that facilitates human checking.
- We are finding limitations, but from that we are gaining useful insights into practitioner usage.
- It is not the job of a judge to do the lawyer’s work for them but understanding these tools and how they are being used informs what the judge does, from timetabling orders to the questions asked of counsel.
Another area the Supreme Court is hoping to explore is the potential for AI to provide fast, reliable answers to process queries from Court users, allowing human effort to be directed to complex queries and the processing of cases as volumes increase. The aim is to begin with a high-volume, low risk area with clear fixed processes as a proof of concept that can then be scaled for more general use.
It is likely that the principal investment in AI will remain with the profession rather than the Courts. In that setting, it is possible that the Courts will be able to adapt their processes to reflect the increased use of AI by the profession. For example, our expectations of what can be done and when, in the context of preparing matters for trial, may need to be recalibrated in the light of any efficiencies that responsible use of AI can yield.
In order to make informed decisions about the use of AI within our Courts, Judicial Officers and Court staff must be familiar with the technology, test what it can do and interrogate various use cases. We have started on that journey.
Benefits
We should measure the success of the adoption of AI in terms of how it has supported the defining features of the profession that I have mentioned: the application of legal knowledge and judgment in an ethical setting. And in relation to the Courts, trust in the administration of the law by human judges must remain critical.
We know there are risks associated with AI use. That said, provided the risks are managed though governance and investment, it is to be hoped that there will also be significant benefits to the profession and your clients. And in relation to the Courts, benefits to the community as a whole.
We have seen very significant technological change in the practice of the law since I was admitted, but it seems doubtful that those changes have resulted in more accessible justice. That pattern should not repeat itself in the case of AI.
When considering the role of AI in our judicial system, our focus should be on tangible improvements for Court users. The use of AI must be seen in a world in which cost and delay threaten to undermine access to justice both in the civil and criminal jurisdictions. And its adoption should be directed to ameliorating these burdens. If AI is to achieve its promised potential, there must be a public dividend and we should examine what public good can be extracted from the technology.
When you do the calculations about risk and reward from investment in AI, I invite you to consider the potential for broader benefits and how the adoption of AI can make the profession stronger. That requires a clear-eyed focus on those attributes of the legal profession that are unique and important in the administration of justice. Equally, when we look at how AI can be adopted by the Courts, the same features must be central.
Conclusion
Public trust in the use of AI is earnt through transparency, understanding and experience. I hope that publication of the Court’s new Practice Note and Judicial Guidelines will assist with this process.
As Chief Justice, I am committed to fostering an environment which facilitates excellence, and I am proud of the dedication, care and expertise of our Judicial Officers and Court staff, as well as their thoughtfulness and adaptability in times of change.
Victorian Courts and Tribunals have a strong culture of embracing digital innovation to great effect, and I remain cautiously optimistic about the opportunities AI can provide to enhance access to justice and the experience of Court users, as well as to support those working in our Courts.
Today’s meeting will be an invaluable exchange of ideas and I look forward to hearing about the present but also about what lies ahead.
Thank you.
[1] National Center for State Courts, State of the State Courts: 2025 public opinion poll findings.
[2] Supreme Court of Victoria, SC Gen 25 – The Use of Artificial Intelligence by Court Users.
[3] Supreme Court of Victoria, Guidelines: The Use of Artificial Intelligence by Judicial Officers.
[4] Australian Trade and Investment Commission, Artificial intelligence and data centres.