Artificial Intelligence and litigation – how’s it being used, what are the risks?   

Written by Daniel Johnston

“To a degree, the future must remain unknown. Artificial intelligence and its effect on Courts, the profession and the law will change the landscape of life in ways we cannot predict.”

Chief Justice of the Federal Court of Australia – The Honourable Justice J Allsop AO

It is the dawn of a new age. AI, like fire was to our distant ancestors, has the potential to reshape how we as humans do things in general, and it will be a game changer to our evolution no doubt. The law, the Court’s, and how the legal profession practices law and how clients interact with their lawyers are no different.

Like our more recent predecessors who grappled with the introduction of the internet and things like instant electronic communication via emails, lawyers now must face and embrace the rise of AI as it has become a daily tool used by millions, at all levels of society.

But how is AI being used in litigation? It is a question perhaps few have thought about unless you are somebody who is at the coal face of litigation every day. It is impossible to limit the categories of how AI is or will be deployed in litigation, but we can at least look at some of the widespread contemporary examples of where it is being deployed to see how it might shape how litigation is practiced in the future and what the risks are in using it.

Some examples of how AI is currently being used in litigation are as follows:  

1. Legal research and e-discovery;
   1. E-discovery software enables large troves of documents to be analyses and classified; and
   2. AI virtual assistants that are designed to help lawyers conduct legal research and provide case and legislation summaries.
2. Document automation;
   1. Streamlines document automation for things like pleadings, letters of demand, and applications.
   2. Storage of large volumes of documents in discovery
3. Predictive legal analysis;  
   1. Maps out litigation trends;
   2. Predictive AI that examines past data to foresee case outcomes.

Some of the obvious benefits here are things like efficiencies, and increased productivity, a reduction in costs. Deploying AI in those circumstances will free up lawyers from mundane tasks that really do not add any value to any client’s matter, like developing stronger arguments, or more persuasive cases or a better case theory. It should permit their lawyer to spend more time on things that actually matter.

Litigation is full of inherent unknowns, and this is obviously and understandably very frustrating for clients (and for lawyers too who are searching for answers for their clients), who crave certainty in the outcomes of their matters, when guarantees cannot be given, in circumstances where, a range of possibilities can happen.

AI tools can help limit some of those outcomes to the most probably or plausible, based on raw data, if they come from a reliable data set. Obviously, this is a very attractive proposition to both lawyers and their clients, but does it also come at a risk? The answer is “yes”. Our distance ancestors who mastered fire, inevitably did so, but not without getting “burnt” a couple times in the process.

So what are some of the risks in using AI in litigation? A full analysis of these risks is beyond the scope of this article, but a few include the following:

1. Hallucinations – which occur when answers generated by AI are factually incorrect or entirely fabricated despite sounding plausible
2. Bias – which can occur when bias answers are generated by AI based on a user’s prompts.  
3. Confidentiality – where is the data being stored? Who has access to the data that is being used in the AI tool?

There are now a growing number of cases, and published decisions in this space that practitioners, and their clients, need to be aware of where that are examples of how this risks have manifested themselves in real matters. [1] These cases outlined that the consequences can be severe if AI is misused, or not respected as the tool that it is, and not the Oracle that some litigants want it to be.

AI is evolving fast, and so too will the risks associated with its use. There are some easy wins though. Many of the Courts throughout Australia in all jurisdictional levels, have issued practice notes regarding the use of AI which contain useful guidance on AI’s use, and the Courts’ expectations on its use. Lawyers, clients, and self-represented litigants should make themselves familiar with these practice notes, especially if they intend to utilize AI in their practice or matters.

Whether it is a lawyer, a client, or a self-represented litigant, the user should be aware of the quality of the tool they are using. For example, the quality of the legal research produced will be influenced by the dataset from which the chatbot is operating. Anyone who is seriously considering using AI in litigation should sensibly use reputable legal publishers that have closed data sets. Most legal publishers now have their own Gen AI tools that are based on their existing data and cases.

So while we have to be clear eyed about the risks of AI, it is likely to become a positive force in litigation. It will likely free up lawyers to pursue more high value work, and can improve efficiency and productivity, so long as it is balanced with trust, fairness, and other things like duties to the Court and clients.

His Honour, Justice J Allsop AO’s words are wise. However, while we may not predict with any degree of real certain how AI will change the landscape of law or litigation, we know change is coming. Like our distance ancestors, and contemporary predecessors, we need to adapt and respect this new, powerful tool, so that we understand its risks, how we can use it responsibly, have guard rails in its use, and know its boundaries, so that the obvious benefits can be derived for lawyers, clients, self-represented litigants, and the Courts. 

 Authors Note: No AI was used in drafting this article.   

[1] See for example Murray on behalf of the Wamba Wemba Native Title Claim Group v Victoria [2025] FCA 731; Valu v Minister for Immigration and Multiculural Affairs (No 2) [2025] FedCFamC2G 95; JNE24 v Minister for Immigration & Citizenship [2025] FedCFamC2G 1314