Want to listen to the full episode and all our other episodes?
Hearsay allows you to fulfill your legal CPD requirements every year.
Our yearly subscription is only $299/year.
With a yearly subscription, you can access all of our episodes AND every episode we release over the next year.
Episode 149
Buy Episode
Barristers’ Skills Ethics and Professional Responsibility Professional Skills
The Devil’s in the Details: Drafting Affidavits That Speak for Themselves
Law as stated: 12 March 2025
What is this?
This episode was published and is accurate as at this date.
Dr. Sonya Willis, Senior Lecturer at Macquarie Law School, joins David to delve into the nuanced skill of affidavit drafting. With direct speech evidence becoming an increasingly hot-button issue following Kane’s Hire, they unpack how to navigate evolving expectations, avoid common pitfalls, and ensure affidavits hold up under scrutiny.
Barristers’ Skills
Ethics and Professional Responsibility Professional Skills| What area(s) of law does this episode consider? | Drafting affidavits. |
| Why is this topic relevant? | Affidavits play a vital role in the courtroom, providing a way for witnesses to share their evidence in a structured, written format. But drafting them isn’t as straightforward as it might seem. In fact, the way affidavits are written has been under the judicial microscope recently, with some strong criticism directed at practices that blur the lines between the witness’ accurate recollection and the lawyer’s own interpretation. Because drafting robust affidavits is a craft in itself. It isn’t just about putting words on a page, it’s about presenting evidence clearly to the court whilst preserving the integrity of the witness’s own recollection in their own words. Missteps here can have serious consequences, from affidavits being rejected to breaches of professional conduct rules. |
| What legislation is considered in this episode? | Evidence Act 1995 (NSW); Evidence Act 2008 (Vic); Evidence Act 2001 (Tas); Evidence Act 2011 (ACT); Evidence (National Uniform Legislation) Act 2011 (NT) (Evidence Act) |
| What cases are considered in this episode? | White v Data Transfer Services Pty Ltd & Ors (No 2) [2022] NSWSC 963 (White v Data Transfer)
Kane’s Hire Pty Ltd v Anderson Aviation Australia Pty Ltd [2023] FCA 381 (Kane’s Hire)
Lantrak Holdings Pty Ltd v Yammine [2023] FCAFC 156 (Lantrak Holdings)
Gan v Xie [2023] NSWCA 163 (Gan v Xie)
Wild v Meduri [2024] NSWCA 230 (Wild v Meduri)
Watson v Foxman (1995) 49 NSWLR 315
|
| What are the main points? |
|
| What are the practical takeaways? |
|
| Show notes | Supreme Court of New South Wales, Practice Note SC Gen 23 – Use of Generative Artificial Intelligence (Gen AI) Head, A & Willis, S 2024 ‘Assessing law students in a GenAI world to create knowledgeable future lawyers’, International Journal of the Legal Profession, 31(3), pp. 293-310 Hack, A & Willis, S 2024 ‘Direct speech evidence after Kane’s Hire and Gan v Xie: did you get the gist?’, Law Society Journal Online Hack, A & Willis, S & Fernandez, S 2024 ‘Wild ride for NSW lawyers: new guidance on direct speech evidence’, Law Society Journal Online |
DT = David Turner; SW = Sonya Willis
| 00:00:00 | DT: | Hello and welcome to Hearsay the Legal Podcast, a CPD podcast that allows Australian lawyers to earn their CPD points on the go and at a time that suits them. I’m your host, David Turner. Hearsay the Legal Podcast is proudly supported by Lext Australia. Lext’s mission is to improve user experiences in the law and legal services, and Hearsay the Legal Podcast is how we’re improving the experience of CPD. In the vast majority of civil proceedings, affidavits are the way that witnesses share their evidence with the court in a structured written format. But drafting them isn’t as straightforward as it might seem. In fact, the way affidavits are written has been under the judicial microscope recently, with some strong criticism directed at common practices that blur the lines between the witness’s accurate recollection and the lawyer’s own interpretation. Drafting a good affidavit is a craft in itself. It isn’t just about putting words on the page, it’s about presenting the evidence clearly to the court in accordance with the rules of evidence, whilst also preserving the integrity of the witness’s own recollection in their own words. And missteps here can have serious consequences, from affidavits being inadmissible to breaches of professional conduct rules. Now, to help us unpack all of these latest rulings and how to best draft affidavits in light of these rulings, we’re thrilled to be joined today by Dr. Sonja Willis, a senior lecturer at Macquarie Law School, my alma mater, and a member of the Law Society of New South Wales Litigation Law and Practice Committee. Sonya, thank you so much for joining me on Hearsay today. |
| 00:01:38 | SW: | Thank you so much for having me. |
| 00:01:40 | DT: | Now, I’m excited to talk about this one. As some of our listeners might know, I used to be a barrister in another life. And so I’m excited to get back to talking about drafting affidavits, very exciting stuff. But before we do that, tell me a little bit about your journey in the law. |
| 00:01:51 | SW: | I’m a bit concerned you have more expertise than I do on this topic, but that’s fine. I’ll just hand over to you. My background is, I worked for a firm that’s now called Ashurst, for nearly a decade as a commercial litigator and dispute resolution person. And then I moved into academia and I did a PhD at Sydney Uni on civil procedure, which is probably something most people would think, why would you do that? And I’m a bit obsessed with procedural fairness. So that’s why I did that. And I’ve written a few books and I teach and research mostly in civil procedure with a little bit of taxation on the side to keep me – something. |
| 00:02:24 | DT: | Yes, I’m not sure what a little bit of tax on the side keeps you really. |
| 00:02:26 | SW: | I’m not sure what it keeps me. |
| 00:02:26 | DT: | Certainly not sane. |
| 00:02:28 | SW: | No, definitely not sane. |
| 00:02:30 | DT: | No. Well, you might say that it’s unpopular to have that much of a focus on civil procedure, but I’ve always found the court rules quite interesting. And I remember very early in my career as a solicitor at a large firm, not too dissimilar to Ashurst, I was told that as a litigator, that’s your armoury. To know – in that team, it was – the Federal Court Rules. And you really had to know those back and forth to know the weapons at your disposal. And I suppose to a degree, the same is true of the Evidence Act. And we’re talking today about how we draft affidavits and some recent, I suppose you could call them controversies, about some of the received wisdom around that topic. We’re talking about how conversations, or what we might otherwise call direct speech, is presented. We’re talking about an oral representation or a recollection of a conversation or a meeting. And there’s been a standard practice around that for a long time, and I think the standard practice that certainly I’ve used for a very long time is, “and so and so said to me, words to the effect of, colon, new line, quotation mark,” and then pretending that we remember exactly what was said with the quotation marks and all. Is that the standard practice? |
| 00:03:32 | SW: | Well, it was, but it arguably isn’t, at least in the Federal Court. I mean, to step back a little bit, you and I are having a conversation now, right? And if you asked me the questions that you’re going to ask me another day, I would probably give you another answer. And that’s true of everyone. So we know that people’s memories aren’t actually perfect and accurate, but it’s a bit like Winston Churchill saying that democracy is the worst form of government except for all the other forms. We don’t have a better way of doing this. So the best way that we have of getting evidence out of people when it’s about things that got said in the past is that you actually ask them. And so normally that would happen in a courtroom, right? You’d have an examination or a cross examination. You’d get the witness in the box and you’d say, “tell me what happened.” And they tell you – well, they wouldn’t make up, they would, to the best of their recollection, give you – a story. And if somewhere in what they’re talking about, they have to tell you about a conversation that took place, you would be hoping that they would be giving you the words of that conversation as best they could recall them. But what they recall today is not going to be exactly what they would have recalled if you’d asked them last week, and it will be different again from what they would say in a month’s time. And that’s where the heart of the controversy is in this case. It’s about what happens when you don’t have exact recollection. So I think the way that we can think about this now – and I have a debt to Justice Jackman from cases like Kane’s Hire that we’re going to talk about – is that there’s three kinds. There’s verbatim evidence; “I remember exactly what was said.” There’s best recollection evidence; “I’m pretty sure I remember that we said something like this,” and then there’s, “I’ve got no idea what anyone said.” And I think really all of the problems are in that middle area, not the outsides. I think Justice Jackman in the Federal Court agrees wholeheartedly with Chief Justice Bell in the Supreme Court in relation to those extremes – the verbatim evidence and the, “I’ve got no idea what anyone said.” It’s that, where I’ve got a best recollection version, that we’ve got some difference now that’s happening. |
| 00:05:27 | DT: | Yes. And you mentioned that there is some difference in the Supreme Court and the Federal Court. Let’s do it chronologically, I suppose. Talk about the Supreme Court case first. That’s White v Data Transfer Services Pty Ltd & Ors (No 2) [2022] NSWSC 963. What was the holding so far as how affidavit evidence of conversations is presented in that case? |
| 00:05:45 | SW: | Well, I don’t really think White v Data Transfer was a particularly significant case. I think it was just a case which drew attention to the fact that there were some issues here. And to be honest, I actually think what triggered this was Kane’s Hire. And I think that the reason it got triggered was because there’s been some very bad lawyering going on. I think actually at the heart of this controversy, if you like, is some very bad lawyering. And the very bad lawyering – and I’m not going to name names or have cases or anything like that, I’m going to use a hypothetical – but let’s say you imagine you have a couple of people that are witnesses that you’re going to need evidence from them. You’ve got some documents that are relevant, and so you think, what would they probably have said? And you make it up and then you put it in, “I said words to the effect of,” and “he said words to the effect of,” and then you give it to your poor witness who knows nothing about law and is relying entirely on you as the lawyer. And you say, “this is your affidavit. You should sign this.” And they go, “okay.” And they sign it and then they end up in the witness box being cross examined, and they get taken to this conversation that was in their affidavit and asked why they said that. And they say, “I never said that. I never would have said that. That’s not my words at all.” Now, that is the sort of thing that Justice Jackman is getting upset about in Kane’s Hire, with very good reason, because that’s actually appalling lawyering and that should never happen. When you’re taking a witness statement, you should be actually interviewing your witness and finding out what they said and what happened, and then that should be recorded as accurately as possible in your affidavit. So I think, at the crux of this controversy, if you like, is really just judges wanting lawyers to do a good job, which is something that I think they all agree on and we all agree on. |
| 00:07:21 | DT: | Well said. And I think it’s a root cause at the root of the symptom, right? We talk about this as, “oh, new case law about how we do conversations in affidavits.” No, this is the professional responsibility upon us when taking instructions on affidavits. |
| 00:07:36 | SW: | Precisely so. And so I think when you look at the cases where the judges have gotten particularly irate, and Kane’s Hire is one example, Lantrak is another. Those are the Federal Court cases that Justice Jackman talked about this predominantly in. That’s really what his beef is. He is upset that people are basically putting words into their clients or witnesses mouths that those witnesses never said and never would have said. And what’s more, they’re happy to stand in a witness box and say, “there’s no way I ever would have said that.” That’s really not good lawyering. That should not be happening. |
| 00:08:11 | DT: | Yeah. |
| 00:08:11 | SW: | But it clearly is or has been. TIP: Now Sonia just mentioned Justice Jackman’s criticisms in Kane’s Hire Pty Ltd v Anderson Aviation Australia Pty Ltd [2023] FCA 381. Since the mid 1990s, lawyers in New South Wales have commonly presented evidence of indirect speech, using quotation marks or phrases like, “the witness said words to the following effect: [… ] ” New South Wales lawyers might be surprised to learn this isn’t a universal practice in Australia. The respondent’s lawyers in Kane’s Hire were from Victoria, and they used indirect speech, presenting the essence of the conversations rather than direct quotations. Justice Jackman criticised the practice of using direct speech when witnesses can only recall the gist of a conversation. Now Justice Jackman emphasised that there’s no legal rule in Australia requiring direct speech for evidence of conversations, though, again, many New South Wales lawyers, advocates, sometimes even judges – well, behave as though there is. He argued that direct speech should only be used when a witness can genuinely recall the exact words spoken. He also criticised the practice of “working up a conversation in direct speech from a witness’s memory of the substance rather than the form of the conversation,” calling it “logically, ethically, and grammatically wrong.” He proposed that evidence should accurately reflect the type of memory that a witness has of a conversation, whether it’s verbatim or a general recollection. He argued that when the witness can only remember the substance of a conversation, not the exact words, it should not be presented as direct speech. He then went on to provide six principles for presenting evidence of a conversation, stressing the need for clarity between exact memory and what we’re calling gist memory. Now in Lantrak Holdings Pty Ltd v Yammine [2023] FCAFC 156, Justice Jackman reaffirmed his stance on the criticisms that he raised in Kane’s Hire, but this time on the bench of the Full Court. Again, he reiterated his criticism of the practice of presenting evidence of conversations in direct speech when the witness can’t recall the exact words. |
| 00:10:06 | DT: | We’re talking about this in the context of recollection of conversations. It is, to a degree, an issue that infects other parts of affidavits. I think maybe judicial officers have been more forgiving of the lawyer’s drafting of other paragraphs in an affidavit that don’t go beyond the recollection of the exact words of a conversation, but from time to time you see a witness cross examined on their use of a particular word or phrase in their affidavit that they don’t actually know the meaning of and judges will say, “well, I know how these things are done. I know that the witness didn’t draft this,” but it’s an acute issue where there are recollections of conversations because often matters that involve evidence of that kind – a fact in dispute probably turns on the words that were used in that conversation. |
| 00:10:56 | SW: | Exactly. And in fact, one area where this becomes incredibly complex and difficult is where you’ve got witnesses who don’t speak English. And so everything is going through an interpretation process, and so their evidence in their affidavit is actually not what they said, because what they said was something in a different language. So everything you’re doing is as close as you can manage it, but if they are talking about a conversation, which for example, changed a contract. This is like a really common example. You have a contract, it’s in writing, and then there’s a variation and that variation takes place orally. Something happens. You didn’t manage to deliver the bananas because they went bad. There was a storm. We can all sympathise with that. And so then you said, “well, can’t deliver the bananas, but maybe I could deliver you mangoes.” And you go, “yeah, mangoes would be great. Your mangoes always walk out the door, take the mangoes instead.” It wasn’t in the contract, but it’s really critical. So then you deliver the mangoes, and then they say, “I’m not going to pay for the mangoes, I ordered bananas.” And that conversation is really critical to the dispute. But maybe that all took place in Chinese, and you’ve just had to translate it. It’s still direct speech, it’s still a conversation. And so if you’re looking at how you’re going to represent that in an affidavit, really, you want to be doing it, ideally, in conversation. At least, I think that that’s what you should be doing and that’s what Chief Justice Bell thinks you should be doing. |
| 00:12:09 | DT: | Yes. Alright, well let’s go into some of these cases and I suppose, what matters is the conclusions of the judges who’ve decided them in terms of how that evidence should be presented. So let’s talk about Kane’s Hire first. That’s a 2023 decision of the Federal Court. What did Justice Jackman, I believe, say about this type of evidence in Kane’s Hire? |
| 00:12:33 | SW: | Well, what he essentially said is that you should only be putting something in inverted commas where your witness is a hundred percent sure of what they said, and this is where we get this verbatim evidence term. So he’s saying, if the witness says to you, “he absolutely said mangoes would be perfect. I remember that phrase. He said mangoes would be perfect.” That should go in quotes, because he’s absolutely certain that that is the exact phrase that was said. But, if he’s talking about the whole conversation about the mangoes and the bananas, and he can’t remember exactly what words were said, then Justice Jackman says that shouldn’t go in inverted commas. That should be gist evidence – I remember the gist of the conversation was this […], and then it should be inverted commas, not expressed as quotes. And he said that because he said, “the witness doesn’t really remember the exact words that were said, they’re going to change their mind about what was said the next month, the next week, so therefore that shouldn’t be in quotes.” And I have some sympathy with where he’s coming from there, but I also think when you’re talking about a conversation, and when you’re talking about an affidavit, if you think about what an affidavit is, what these affidavits are essentially, is evidence in chief. They are a written version of evidence in chief. So they’re a way that you get your witness in the box without getting them in the box. And they’re fantastic for complex commercial litigation because often the facts and the evidence are so complicated that you’d be there for days and days trying to sort it out, and it’s so much better if the lawyer can sort it out and have it all set out nicely in an affidavit and then you can be cross examined on it. So one of the things that’s gone wrong, perhaps, is that people haven’t been thinking about affidavits in that way as if they are really a version of evidence in chief, and I guess that’s my objection to Justice Jackman’s interpretation, is that I feel like he’s taking it a step further away from being evidence in chief in a written document. |
| 00:14:26 | DT: | Yeah, it’s a useful heuristic, isn’t it, to think about it as a substitute for oral evidence in chief? Because as a commercial litigator, you yourself, with your own experience in that field, you seldom ever see oral evidence in chief, except perhaps in an interlocutory hearing where no one’s had time to do the affidavit. |
| 00:14:42 | SW: | Yeah. |
| 00:14:42 | DT: | But as a university student, I was a court reporter in the criminal courts, in the Parramatta District Court, where of course all evidence is oral, in chief or otherwise, and I suppose people have very different idiosyncratic views on how they might give an account of a conversation. Some people might say, “and then I said,” “and he said,” “and then I said,” “and then she said,” some people might say, “well, I said I didn’t want the mangoes,” but not using those words. So it really does help, I think, and perhaps it goes in some cases in favour of Justice Jackman’s interpretation, in some cases in favour of Justice Bell’s, to think about it as, well, this is the oral evidence written down. The Full Federal court weighed in on this as well, didn’t they? |
| 00:15:23 | SW: | Well, in Lantrak, it was really just a majority judgment of which the person that was really talking about it was Justice Jackman, again, confirming what he had said in Kane’s Hire. So essentially, that added an extra seniority level of approval for it. And I think where things got complicated was that in the Supreme Court in Gan v Xie, they agreed. They said, “yes, we think Justice Jackman got it right.” And then of course, Chief Justice Bell came along with the court in Wild v Meduri and said, “no, I don’t think so. I think New South Wales does it the way New South Wales does it.” TIP: Now, Justice Jackman’s criticisms of the practice of using direct speech have also been applied in some New South Wales decisions. In Gan v Xie [2023] NSWCA 163, that court followed Justice Jackman’s reasoning from Lantrak and Kane’s Hire and said that a witness’s inability to recall exact words doesn’t automatically render their testimony unreliable, especially when they say that they can recall the substance or gist of the conversation. And the Court of Appeal confirmed that affidavit evidence, even when presented not in direct speech, is admissible when the substance of the conversation is clear from the affidavit. However, just the next year, the Court of Appeal took a different approach in Wild v Meduri [2024] NSWCA 230. In that case, Chief Justice Bell and Justice of Appeal Kirk criticised Justice Jackman’s position, defending the long standing New South Wales practice of using phrases like, “to the following effect”. They said that this practice allows witnesses to present the gist of conversations and then that can be tested under cross examination. I suppose the argument they’re making there is that no one reading the affidavit in a New South Wales court is proceeding on the basis that that is evidence of the exact words. Chief Justice Bell argued that it would be undesirable for a single judge to unilaterally end such a long standing practice. Judge of Appeal White also made an interesting point. He refrained from directly criticising Jackman’s approach but said that recollections of the effect of a conversation might be treated as lay opinion evidence subject to the rules of admissibility. |
| 00:17:24 | DT: | But it was really Kane’s Hire that was like, the case that launched a thousand LinkedIn posts, wasn’t it? Because this was the one that really challenged the received wisdom that we’ve had for so long that is passed down from partner to graduate – certainly the instruction I got and the instruction I gave on how this is done – and I suppose outside of New South Wales, we have listeners in all jurisdictions in Australia, there really is no inconsistent authority for that, is there? |
| 00:17:50 | SW: | Well, I think actually the interstate thing is a really interesting issue because in fact in Victoria they don’t use direct speech as much and so that’s, I think, an argument in favour of Justice Jackman, his approach is, “let’s be a little bit more consistent,” and I have to say, I think the only reason why personally I’m really sympathetic with Chief Justice Bell’s opinion expressed in Wild v Meduri is that I think it is optimal. I think it’s right, and I don’t really understand what Justice Jackman’s objection to “words to the effect of” is as a phrase, because I had never really thought it was a particularly complicated word. Like it’s, “I said words that had this effect. I can’t remember if these were the exact words, but if they weren’t, there were some other words which had this effect.” So to me, ‘words to the effect of’ wasn’t particularly formulaic. Yes. Okay, probably we don’t normally walk around saying, “I said words to the effect of,” though I sometimes use it just for fun, but probably most people don’t. But you could say “like.” That is what we actually use. That is what people use, and if you’re a teenager, interminably, people use the expression ‘like.’ “I said something like.” And that is really exactly what words to the effect of means; “I said something like ‘X’.” “I said something like, ‘give me the mangoes instead,” or, “I said something like,’ I’m really sorry about the bananas. They fell off the back of the truck’” – or whatever happened to them – “‘they were destroyed in the storm.’” I personally think there’s a place for best recollection evidence because I think it’s the best alternative that we have, like the democracy being the least worst form of government, right? It’s not perfect. We know people’s memories aren’t perfect, but what I think the value of, when it comes to interviewing witnesses, which I did a lot, is that let’s say a witness says to you, “oh, then we varied the contract. We changed it.” That’s very conclusory. |
| 00:19:39 | DT: | Well, and probably inadmissible in that it draws a legal conclusion. |
| 00:19:41 | SW: | Exactly. And this is the issue, right, is that you start to get evidence that’s so conclusory because that’s how people tend to think, and one of the jobs that lawyers have in drafting an affidavit is to try and get the material that they know needs to be before the judge into admissible form. That is really what you’re trying to do, and one of the ways to make a very conclusory statement admissible is to actually drill into it and say, “well, what actually happened? What did you actually say?” And so I think it’s frustrating if we’re in a situation where there’s been a conversation. That conversation is really important. That conversation has caused the contract to be varied or has caused whatever it is, the grief, the beef, you know, maybe it’s a neighbourhood dispute and people are fighting with each other – “she said this,” “he said that.” The “she said,” “he said” is really important. If you say, “well, if you can’t remember exactly what was said, then I don’t want to know what you think was said.” I think that’s really cutting off your nose to spite your face, in a way. That’s really limiting your options, and it’s going to be frustrating for witnesses if you interview them and ask them what happened and what was said and get all this conversation from them and then say, “well, we can’t actually use it because you’re not a hundred percent certain.” |
| 00:20:49 | DT: | Yeah, that’s right, but it’s an important point around the admissibility, isn’t it? That a paragraph in an affidavit that said, ‘and then we changed the contract’ is inadmissible as a legal conclusion. But a paragraph that says, ‘and then I said, “let’s change the contract”’ is admissible, for the truth that it was said, not necessarily for the truth of the statement. |
| 00:21:07 | SW: | And of course, that’s not what you’re going to say. It’s not going to be, ‘we changed the contract.’ That’s going to be, ‘I said, “can I do this instead?” and he said, “yeah.”’ And it won’t be, ‘I said’, ‘he said,’ because that would be verbatim evidence. And your witness has said to you, “I can’t remember exactly.” And in fact, my view, and I think this is absolutely what the Chief Justice is advocating in Wild v Meduri for New South Wales, is that you would put that in the affidavit. You would actually put in the affidavit, ‘we had a conversation and I can’t remember exactly what I said, but I said something like […]’ and he said something like […].’ If that’s the words that your witness gives it to you in, then that’s probably what should go in the affidavit, because if they, going back to our point, were in the box giving evidence in chief, that’s what they’d be saying. The barrister would be saying, “so tell us what happened.” And they’d be saying, “well, […].” |
| 00:21:55 | DT: | That’s right, and I suppose the other point about that is, This does depend to a degree on the extent to which the exact words are important, right? |
| 00:22:06 | SW: | Precisely. |
| 00:22:07 | DT: | There are cases in which evidence of a conversation will be given, but nothing turns on that conversation, nothing turns on the exact words. And I would also expect – witnesses aren’t stupid – a witness giving evidence in an oral examination in chief would be more precise in the language to describe that conversation if they were aware that so much turned on that, so far as the findings of fact, compared to something where it was a fairly satellite issue. |
| 00:22:34 | SW: | Yeah. And I think one of the things – this is another good, “how do you draft an affidavit?” tip – is you should only be putting in things that are relevant. That is actually, when I think about, what should lawyers be doing when they’re drafting affidavits? The most important thing they should be doing is working out what is relevant and what is the best way to communicate that, the most efficient, effective way to communicate that to a court? That’s your job as a lawyer. It’s not to make stuff up that your witness might have said, but didn’t, or that your witness says, “there’s no way I would have said […].” Please don’t do that, but rather, to say, “well, I’ve sat down, I’ve interviewed them, we’ve all done this, the witness goes on and on.” The better you get them talking, the more of the story you get. So you really want them to be in free flow and telling you everything. And they do that. They don’t tell you it in chronological order, they don’t tell you it in admissible form. That’s your job. So your job is to take everything that they’ve given you and to work out which of this is relevant. Where you can see there’s something relevant and they’re not really getting to it, that’s where you start asking them the questions that elicit that information that you know is relevant that they don’t, because clients and witnesses by and large, the sort of people that end up in these affidavits, they don’t know the law usually, and so they don’t actually understand what the legal issues are. So they don’t really know what’s relevant. That’s why self represented litigants do such an appalling job of drafting their affidavits. |
| 00:23:53 | DT: | Yeah. Well, they certainly tend to err on the side of inclusion. |
| 00:23:57 | SW: | They certainly do. |
| 00:23:57 | DT: | To put it mildly. I mean, this comes back to what we were saying before. The lesson about these cases is different to the one that LinkedIn posts are talking about, right? I think when Kane’s Hire was decided, especially lawyers in New South Wales said, “oh, look, completely different approach, different to what we usually do, do we all have to change the way we draft affidavits now?” Where we’ve sort of ended up is back with the status quo. There’s some difference of opinion in other jurisdictions. New South Wales still do it the way they do it, as you say. |
| 00:24:28 | SW: | Well, except that we are, of course, often appearing in Federal Court cases. And that I think is the challenging area is what happens if you’re a New South Wales lawyer who runs cases in both the Supreme Court and the Federal Court. Do you have to take a different approach to drafting depending on which court you’re appearing in this week? And the answer is probably for a little bit, yes. To a certain extent, yes. But like you say, we’re only talking about where that conversation is really important, and we’re only talking about where that conversation is one where the witness can tell you in an actual conversation words what happened, but they’re not a hundred percent sure of the recollection. That’s really the only point of difference. So I think it is important not to panic. |
| 00:25:07 | DT: | Yes. This is your middle category of best recollection evidence. Yeah. |
| 00:25:11 | SW: | Best recollection. Yes, that’s what I’m calling it. Yeah. Best recollection. So yeah, the verbatim and gist are terms that Justice Jackman has given us. Best recollection comes from Chief Justice Bell. I’m not making any of this up, I’m stealing from everyone. So yeah, I think that’s where you get your three categories. Verbatim, we’re all in agreement, everyone says you should put it in quotes; gist, we’re all in agreement, client doesn’t remember what was said, can’t give you the wording, definitely shouldn’t be in direct speech in anybody’s book. But it’s that middle category where they really are able to give you the words, but they’re not a hundred percent confident. And maybe you ask them one day and then when you’re going back over their evidence to finalise the affidavit a week later, they say, “oh, actually, I think it was this.” Do you at that point, pull it out and say, “well, we’re not going to have this conversation.” And the answer there is, if you’re in the Federal Court, you pull it out, and in fact, you shouldn’t have had it in the first place post Justice Jackman. If you’re in the New South Wales Supreme Court or District or Local, then you apply the Wild v Meduri precedent from the Court of Appeal, and you do put that in your quotation marks with a phrase, and let’s not use ‘words to the effect of’ because you’re going to offend our friends at the Federal Court if nowhere else, so maybe you use what they actually said. What did they actually say when they were relating this conversation to you? If they actually said, “I said something like […]” put that in. |
| 00:26:28 | DT: | I’m actually quite attracted to the idea of using, “something like.” |
| 00:26:32 | SW: | Yeah, well that’s what people say. They don’t say, “words to the effect of.” But it’s exactly the same thing. “Words to the effect of…” and “I said something like…” are exactly the same. |
| 00:26:39 | DT: | It adds verisimilitude. |
| 00:26:41 | SW: | It does. Well, it’s real. That’s right. I love it. |
| 00:26:43 | DT: | I suppose, the lesson underneath all of this as well, which we should all always have been doing, but it’s a good reminder, is you certainly shouldn’t be making any of this up and you should be taking instructions on each and every one of these conversations first before presenting the draft affidavit to your witness. This practice was so commonplace in New South Wales that occasionally objections would be founded on it. On occasion, and I’m sure it still happens, paragraphs of an affidavit would be objected to on the basis that they were not in direct speech, and sometimes those objections would be upheld. I mean, even in New South Wales, some judges will quite rightly say, “well, where in the Evidence Act does it say that?” But do you think that objection is less likely to be taken or upheld now that we have that difference of opinion between the courts? |
| 00:27:24 | SW: | I think so, because I think even though Justice Bell has come back very strongly, in fact, it’s quite astonishing to-ing and fro-ing between those judges. Very unusual to have people getting, I would say, stroppy with each other in writing. |
| 00:27:40 | DT: | What a rare exception to the principle of comity, to deal with drafting affidavits. |
| 00:27:44 | SW: | Yeah, really, really quite shocking. But I think the frustration is understandable. Are judges going to take objection? I mean, obviously, I don’t know, because I am not prescient, sadly. But what I can say is, I think even Chief Justice Bell, he’s talking about the value of best recollection evidence, and he talks about things being conclusory when they’re not in that format. And so I think you’re not going to get objections to evidence on the basis that you didn’t use direct speech, except where that direct speech is really important, and where, as you say, something turns on it. You have a situation where, well, did you vary this contract or didn’t you? And maybe the wording of what you actually said is going to be the driving force in making that decision, and then if you have put this very conclusory rolled up statement in, that’s just not going to be evidence. |
| 00:28:36 | DT: | Yeah. And to your point earlier of, “well, this is less wide ranging of a divergence than you might be tempted to think,” I suppose the other thing we need to overlay over this is that uncorroborated evidence of conversations is always going to be received, if at all, with a great deal of skepticism to the extent it isn’t corroborated by some contemporaneous document, you know, the old Watson v Foxman approach to uncorroborated evidence of a conversation. TIP: Now, I just mentioned the case of Watson v Foxman (1995) 49 NSWLR 315. The then Chief Judge in Equity, Justice McLelland, delivered the judgment in that one. The case is an oft cited judicial acknowledgement of the inherent fallibility of human memory, especially as time passes and disputes and litigation intervene. Justice McLelland noted that memory is not a precise reproduction of past events. It’s often shaped by subconscious perceptions, self interest, post event considerations that are really influenced by the litigation that then intervenes. Justice McLelland emphasised that memory fallibility intensifies over time, especially in the context of civil disputes. As people reconstruct events, their recollections are influenced by subjective interpretations and motivations. I’m going to quote directly from Justice McLelland’s judgment now: That leads me to, I guess, a provocative question, which is, we’ve been saying, “well, tut tut, how terrible it is, lawyers drafting these affidavits first and then seeking instructions on them,” which I agree is bad practice, but to be devil’s advocate here, we know from Watson v Foxman that really what the court’s going to take special attention of is the contemporaneous records of that conversation, more so than anyone’s particular recollection of the conversation itself. I’ve got a file note or I’ve got an email sent five minutes after the conversation. It uses particular language that’s been sent to me. I draft the affidavit and I say, “well, what’s likely to have happened is what’s in the email that happened five minutes after the conversation. And he said this phrase, so it’s probably that phrase. In fact, anything that he says inconsistent with that email is probably not true and is probably not likely to be accepted to the extent it’s inconsistent with the email.” So shouldn’t I just draft it based on the contemporaneous record and then say, “remember the contemporaneous record? That is what you said, right?” |
| 00:31:18 | SW: | Look, and I think this is exactly, you’re getting to the crux of the complexity and the difficulty right to the core of my old PhD thesis that will never go away, which is that there’s this constant balancing act between, for New South Wales people, just quick and cheap, that you’re trying to balance efficiency because you do want to be proportionate in your costs as a lawyer, et cetera, with the obligation you have to get things right and where does that balance lie? And I think essentially what this debate is telling us is that you probably should ring your client and just say, “I’m just wondering, I’m drafting your affidavit. You said this stuff in the correspondence. Is that pretty much what you said?” And away you go. The other thing is where you have contemporaneous notes, as you say, they’re going to be more valuable. So maybe you say to yourself, this isn’t a case where the actual conversation is going to be what the judges are going to be deciding on. They’re going to be using that contemporaneous written evidence primarily to make the decision, and so really what was actually said is less important and maybe that’s a point where the gist evidence is perfectly appropriate. “We had a conversation. I wrote up the conversation in my notes, which is the most accurate representation of the conversation that I have, and it’s annexed at A.” And then that’s all you need, because you’ve got that contemporaneous document. I think as well, David, one of the really bad things that’s been happening and it’s a really interesting point on that recollection is where multiple people are referring to the same document and you’ve got lawyers writing, “words to the effect of,” etc type discussions and it’s identical for multiple witnesses. And we know it’s one of those funny rules where even though we want to recall as best as possible, in fact, when the corroboration is so perfect that everyone said exactly the same thing, the judge becomes very suspicious that maybe people didn’t actually say exactly that at all, because we know that recollection doesn’t work that way. So paradoxically, actually, you’re looking for a little bit of difference between witnesses because that’s the best evidence. And I mean, I’ve done these cases before. I’ve had cases where I had to interview every single member of the board about a meeting that took place and I’ve got the minutes of the meeting and you can’t just say “now,” – you can’t do that. |
| 00:33:39 | DT: | Yeah. Paragraph 20 copied and pasted into every board member’s affidavit. |
| 00:33:43 | SW: | Yes, yes, you can’t do that. Yeah, it does come back to that point of what good practice is and you’re right, it’s really hard when we’re pressed for time and we’re trying to comply with our obligations under section 60 of the Civil Procedure Act to keep our costs nice and proportionate, and minimal, and et cetera. But there are certain corners that you can’t cut, and I think the message that we’re getting from these cases is that’s one of the corners you can’t cut. You can’t just write that affidavit up and say, “I wrote this based on what I thought, based on the documents.” You might be going to say, “is that okay?” |
| 00:34:13 | DT: | Yeah. But this comes back to your earlier point, which was, what is truly relevant to the facts in dispute? What is the best evidence you can put before the court? Not the most, and – although, we’ve all been guilty of that. |
| 00:34:25 | SW: | Yes, it goes back to our self represented litigants again, doesn’t it? The most evidence. |
| 00:34:29 | DT: | The most evidence you can put before the court, if not the best. |
| 00:34:31 | SW: | I love it. |
| 00:34:32 | DT: | You know, I like this idea of, although I’ll regret not having the chance to write “words to the effect of,” maybe I can just go around that entirely. “We had a conversation, I sent an email to Ms. Smith summarising that conversation, it’s annexed.” That’s a great third path through all of this, isn’t it? But we were talking before about what, especially New South Wales lawyers, should be doing who have this, as you say, quite strong opinion from Justice Bell on how things should be done in that court and an uncontradicted opinion in the Federal Court about how that should be done from Justice Jackman. And we do have some areas of the law, where we really have a choice of forum to a degree – but in fact, in the ultimate way, things turn out, maybe not have a choice – I’m talking about things like litigation under the corporations act and solvency matters, the regulation of liquidators and voluntary administrators and things like that, where you really do have your choice between the Supreme Court of your jurisdiction or the Federal Court, the law that applies, the facts in issue are all going to be identical in these cases, but the courts will take quite different views of how that evidence might best be presented. In those cases, I suppose there’s two limbs to this. One, are there any practitioners so willing to die on this hill that they might choose which court they choose to litigate these matters in based on how they can present this evidence, but that might actually be important? |
| 00:35:56 | SW: | Oh yes, I can completely imagine a case actually, now that you mention it, where there’s a conversation that’s really looking very critical and you say to yourself, “well I reckon there’s a trade practice consumer law angle to this, I’ll bring it in the Federal Court because then I know I don’t have to put all that stuff into direct speech. In fact, I’d be wrong to do that, and so I can manipulate this evidence in a way that I would like to.” I’d like to think that wouldn’t happen, but I can completely imagine a scenario, yeah, where strategically it’s another factor in that decision that you make between whether you run this matter in the Federal Court or the Supreme Court, where you’re sort of going, “well, I think based on what my witness is like, I think this evidence is going to look a lot better in a Federal Court – gist, no verbatim, so don’t worry about that, no conversations evidence – than it would in a supreme court where I might feel compelled to have a stab at discussing these important conversations that took place, even though, you know, I know my witness’s recollection isn’t perfect.” |
| 00:36:56 | DT: | Or even thinking about the cases to do with valid appointments of voluntary administrators where directors have to resolve that the company is reasonably likely to be insolvent at some future time. They actually have to turn their mind to that in the meeting. They can’t just say, “let’s appoint a voluntary administrator.” The meetings, from time to time – I can think of a case where it was, the meeting took place at a pub. |
| 00:37:17 | SW: | Standard corporate governance. |
| 00:37:18 | DT: | Yeah, that’s right – how did that company end up in insolvency? – where you might, depending on whether or not you think it will be more or less persuasive to give evidence of the means by which the directors turned their minds to the solvency of the company, choose to do that in direct speech or choose to do that in a little more conclusory, gist type, form. I can see it having a non zero impact on the decision. |
| 00:37:43 | SW: | I can too actually, which is a bit terrifying, but yeah, I can too. |
| 00:37:46 | DT: | But then you could also have the circumstance where matters are transferred between those courts. |
| 00:37:51 | SW: | Yeah. So what are you going to do if you’ve got a case where there’s some insolvency or something involved and you can really see that possibility – you’re in the Supreme Court – but you can see that possibility of a transfer, what do you do? Well, there’s nothing in the Wild v Meduri decision that says, “you must put everything into direct speech wherever possible.” So I think you would go for something where you, as best as possible, avoiding conclusory as much as possible, used as little direct speech as possible. So that’s what I’d be going for. |
| 00:38:23 | DT: | Just walk the tightrope. |
| 00:38:23 | SW: | Just walk the tightrope. Try and deliver your evidence in a way that is going to be happy for both courts, which a lot of the time you’re going to be able to do. And I think part of this as well is just making sure when you’re talking to a witness that you do know what was actually said and how confident they are about what was actually said. And then the other thing is, let’s say you put something in as direct speech, the verbatim evidence thing, and then your witness says something ever so slightly different, cross examined in the box. I don’t think that’s going to fall foul of Justice Jackman’s desired perfection in witness affidavits and analysis and cross examination. It’s going to be a situation where we’re only human. Everyone’s only human. And I think even where your witness swears that they’re absolutely certain of the exact wording, if they come up with a slightly different wording in a cross examination, I don’t think anything massive is going to turn on that. |
| 00:39:22 | DT: | Yeah. Just as a practice point, I think about how I would have conducted those interviews or how I suppose, but for these cases, I probably still would. I’d usually say, “well, then what did you say?” Right? And then often the answer would be, “oh, I don’t remember.” And what I would usually then say to the witness at that point is, “don’t worry too much about the exact words. What I’m going to do is say that it’s words to the effect of what you’re telling me, but the court will know. And we’re going to say in your affidavit that it’s not exact. I just need to understand as best as you understand it, what you said.” But I think what I’ve picked up from our conversation, which I’m going to take forward is, well, before I asked that question and “what did you say?” And then give the council, “well, don’t worry too much. I’m going to say it’s something like this.” You have to make the judgment of, yeah, which of the three categories does it fall into? How good is your recollection here? And if the first thing they said was, “oh, well, I don’t know.” That actually directs you to, “well, I’m not going to say words to the effect of, I’m going to just give the gist then,” because you really don’t have a strong impression of anything close. |
| 00:40:27 | SW: | Yeah, I think that’s such a good point and I think that is where practice might change for the better, is that we actually are starting to talk to our witnesses and think more carefully about what their level of knowledge is and reflect that accurately in their affidavit. Where I think the move with the Federal Court might be a bad thing is I worry that lawyers are going to stop trying to get that conversation. So, where the witness just gives them some conclusory thing, they’re going to go, “oh, that’s okay,” because they would no longer see the value in trying to work out what was said. And I know I certainly had the experience in practice where, because we were like you, trying to put everything in direct speech, wherever possible, the witness would say something witnesses don’t speak in evidentially admissible terms naturally, understandably. And so you would be saying to them, “can you tell me though, what happened?” And so you’re, you’re drilling them if you like, you’re not cross examining them because you’re not leading them, but you are, you’re making them think. And often that making them think I found was so helpful. I would often find that witnesses when pushed to recollect, actually thought about what happened and they go, “oh, no, actually that can’t have been then because actually I remember in that conversation I said X and if I’d said X then things would, no, actually now that I think about it, that conversation must’ve happened the week later,” and such and they’re like, “oh,” and so often I find, I used to find at any rate, when I was interviewing real life witnesses, that pushing them a bit about their recollection and getting them to try wherever possible to give you as much detail, including, for example, what was actually said, really helped them to think carefully and nut things out, and I often felt you got much better evidence as a result. You got a much more accurate reflection of what had actually happened, because they’d had to really think hard about it. And when they thought hard about it, they remembered things and they were able to piece stuff together. |
| 00:42:15 | DT: | It sort of impresses the importance of really getting it right, doesn’t it? If you’re that focused on the detail. And this is an example of a broader problem with affidavit drafting, which is just the disconnect between how lawyers tend to speak and tend to write and how many of our witnesses would tend to speak or write in the same context. We were talking about verisimilitude before and I think one of the benefits of really pressing for the exact words is that there can be a credit benefit to that where a judge or some other tribunal effect maybe might notice that, “oh, well, the speech that that witness is giving evidence of, there’s a way of speaking, having heard this person in the box for two hours, sounds pretty plausible. Like that really sounds like them.” |
| 00:43:03 | SW: | Yes. |
| 00:43:03 | DT: | And I wonder if we’re missing out on the same benefit, not taking the same approach to the whole thing? |
| 00:43:10 | SW: | Yeah. Yeah, no, I think that’s entirely possible. And I also think, I mean, this comes back to the, how do you get that evidence in the first place? How do you take that evidence? And, and I said to you before, you push your witness, but you don’t cross examine them. I always used to say, “you do have to cross examine them because you have to work out what, what the cross examiner is going to get out of them so that you make sure that their witness statement is complete.” But of course you don’t lead them, and so that’s the thing. You push them, but you don’t lead them. And I think there’s that important difference, and I think that’s another thing that Justice Jackman was very justifiably upset about in Kane’s Hire and Lantrak, was the suggestion that witnesses were being pushed to say what suited the case rather than what they actually thought, and that’s exactly the kind of leading that you absolutely can’t do. And that’s what you can’t do as a lawyer when you’re taking witness evidence, because it is evidence in chief. So you can’t be saying, “but didn’t you actually say this?”, when that’s what you want them to say. You can say, “might you have actually said this?”, when it’s not what you want them to say, because that’s what the barrister in cross examination is going to put them in. You need to know what they’re going to say in that situation. So you need to actually cross examine them, but not lead them. So you ask the sorts of things that are going to make their evidence come unstuck, because if they’re going to come unstuck, you want to know now. You want to know right at the beginning so that you can settle this case or whatever, you know, or withdraw or whatever, depending on how bad it is. You don’t want your witness to seem fabulous in their affidavit and then come completely unstuck in the box because in fact they were lying to you. So you want to make sure they weren’t lying to you. So, it is quite a tricky business and one that we’re not well trained on. It’s quite a tricky business actually doing a good job of taking witness evidence. |
| 00:44:49 | DT: | Yeah, it’s a bit of a blank space, isn’t it? We’re fairly well trained in the drafting bit for those of us who choose to pursue it, we’re fairly well trained in the oral advocacy bit at the end and cross examining on the affidavit, but the getting the affidavit evidence out of the witness and onto the page is the step in the procedure that we’re left to work out. The kind of, well, this is the soft skills bit, right? |
| 00:45:14 | SW: | And it’s not always done well. That’s exactly right. And I’ve seen situations where people, they’ll say, go in with a list of questions to ask the witness. And I think that’s a really good idea to go in with a list of questions that you plan to ask the witness, and absolutely you should make sure at the end of your interview that you have asked all those questions or that you have the answers to all those questions. But that’s very different from walking in – and I’ve seen junior lawyers do this and it’s painful – and just ask the question and then you ask the next question ignoring whether the witness actually had already answered question six in question three because you’re just following a script and you’re not thinking. And aside from the fact that it’s painful and awful and the witness ends up thinking that the lawyers are stupid, also, the problem with that is that you don’t get everything. Because when you walk in at the beginning of that interview, you don’t know the witness’s evidence. That’s why you’re there, to get their evidence. And so I feel like what should come out at the end when you finish that witness interview is a whole lot of stuff that you didn’t know you were going to find out. That’s a good witness interview, is one where you actually learnt things from the witness about your case. I ran a case once where we had done discovery and we were doing witness interviews and the witness says, “oh, well, actually, I think I’ve got a box of documents in the car.” “Did those go into the discovery?” “Oh, no, I didn’t think of those.” “Oh, well, there’s an amended discovery for you.” You don’t want that to happen, but you would much prefer to find out about that box of documents when you’re preparing that witness affidavit, and then there’s still time to amend the discovery and all the rest of it, than when they’re being cross examined and they say, “oh, back at the office, I’ve got a box of documents.” |
| 00:46:50 | DT: | Yeah, “can I get a production of those?” You’re right, though. You do have to have that more freeform, open approach to the task, because I think it is sensible and you would go to a witness interview saying, “well, what facts in dispute can this person give evidence about? What are the known knowns that I need this person to say so that I can responsibly write it in an affidavit? And what are the known unknowns that I need to investigate? I’m aware that this person will know this thing and that this is relevant, but I’m not exactly sure what happened.” But there are so many unknown unknowns that you should be getting out of that. It’s folly to think you’ve got this perfect picture of the case by the time you sit down with a witness. So yeah, you can’t take that really structured approach. |
| 00:47:30 | SW: | No. |
| 00:47:31 | DT: | You’d have to let them tell the narrative. |
| 00:47:33 | SW: | That’s right. And that comes back to that point I was raising earlier about what is important in drafting is you need to let them tell the narrative and then your job is to say, “okay, they’ve given me this fantastic story with everything they know. I’ve learned all these important things. But I can’t just deliver a recording of my three hour interview because I’ve got to put together an affidavit that is logical, that is preferably chronological, or at least is in an order that makes sense with the case and the issues in the case.” And it needs to be as efficient as possible because judges are short of time. Everyone’s short of time. And that’s the whole value of having an affidavit, having the evidence in chief in an affidavit rather than orally in the box, is that you can get rid of all of that irrelevant stuff and you can think about a structure that conveys the message as efficiently and effectively as possible, and that’s where the other part of the skill comes in. So there’s getting the story out of your witness, putting them sufficiently at ease and making them sufficiently comfortable that they can tell you everything that you need to know, the knowns, and the unknown unknowns, and then also being able to then convert that into something that is as efficient and as beautiful as possible for the judge and the other parties and everyone that’s going to have to look at it to read. Yeah. |
| 00:48:50 | DT: | Now, it’s 2025, which means we, I think, are somehow legally obligated to talk about AI. And speaking of which, we are legally obligated, at least in the New South Wales Supreme Court, to talk about AI in all of our affidavits, because Practice Note SC Gen 23 on the use of generative artificial intelligence comes into effect very soon. It was published last year, another LinkedIn-post heavy publication. The Practice Note says that generative AI is not to be used at any stage drafting affidavits and has a very limited role in drafting expert evidence, which we’ll come onto, but also that every affidavit now needs to contain a statement that generative AI was not used to draft it, I imagine sitting just above the juror, along with the oath or affirmation. Sonya, thoughts? |
| 00:49:40 | SW: | So, look, I think it’s fantastic that the court is doing its best to regulate AI because they recognised that generative AI, or even AI, is everywhere. And I think it’s really important what they are trying to do because if we think about what we’ve been talking about, how critical it is in affidavits to actually get the story, the truth, we can’t think of anything further than asking a brainless piece of computer equipment to make it up for you. That’s got to be the antithesis of good evidence. |
| 00:50:11 | DT: | That’s right. It’s not the witnesses words or the words. It’s just the next token predictor. |
| 00:50:18 | SW: | At least you feel with the lawyers, going back to your scenario – |
| 00:50:21 | DT: | It’s a human being’s words. |
| 00:50:23 | SW: | Well, yeah, and going back to your scenario where you’ve got some contemporaneous document and you’re doing your best, at least that lawyer is trying really hard, you hope, to think, empathically, what would the witness have said? And put that into communication. As we said, we don’t think that’s the way you should be doing things, but it’s a lot better than, like you say, some token predictors coming up with the next most likely word based on what’s in its trillions of data training base that probably has a lot of Judge Judy in it, and goodness knows what other American shows that it’s based in. |
| 00:50:53 | DT: | This was the evidence given on Matlock. Yeah, I think it’s great. Obviously, if there’s a form of legal drafting where generative AI has no place, it is in the document that is supposed to be a true recollection of a real human being. So that’s good. The Practice Note also speaks to the use of it in submissions and in other documents, and I think sometimes, even if we can use it, we’re sort of using it for the wrong thing sometimes. And affidavits have a role to play in being persuasive as well, and this I’m sure will change, but large language models are good at generating text quickly, but it’s not to me, the most persuasive or unique or striking text. I think, even if you could draft an affidavit with the assistance of generative AI, I wonder if viewing it as, “how quickly can I get to the word limit?” or, “how quickly can I get to the page limit for my submissions?” is missing an opportunity to express your thoughts on the case and your understanding of the evidence in a more profound and persuasive way than a large language model might. I actually wrote something about this at the end of last year when the Practice Note was published, but before it came into force, about the decision whether or not to use generative AI for drafting something like a set of submissions not really being a question of whether it was capable of doing the task. If it were capable of doing as good a job at drafting the written submissions as me, or even better, I would probably not choose to do it. And that’s because the intellectual task – and the same is true of affidavits – the intellectual task of reviewing the evidence, reviewing the case law, reviewing the pleadings, understanding your case theory, and distilling it all into this document, either in the affidavit or in the submissions, is a process of understanding your case that you’ll miss out on if it can be automated. |
| 00:52:40 | SW: | Yes. And which, as a barrister, you’re going to need to actually run the case, and so you’re going to spend the time anyway. And look, and this is, I think where it does start to get difficult. Going back to our issue again of trying to be as efficient as possible. Like, are we going to get to a point sometime down the track where people say, “well, your costs were excessive because you did everything yourself when you should have been using generative AI to help you save time.” I hope we’re not about to get there, but I don’t think it’s inconceivable. |
| 00:53:06 | DT: | Oh, it’s certainly not inconceivable. When we think about the obligation that other Practice Notes, including in the Federal Court, impose on the use of eDiscovery tools, especially tools that, say, limit the bundle for human review through technology assisted review at an early stage. Yeah, we are now talking about a professional obligation to use artificial intelligence tools there. |
| 00:53:27 | SW: | So I wrote an article with a wonderful colleague, Amanda Head, last year that talked about this in terms of teaching students. And we decided that, well, students have this obligation to know the law, which is actually legislated, and that means that you have to be careful about the use of AI in testing them because you have to make sure they know it and not that some computer can come up with something similar. But also, we decided they needed to understand AI because of all the ethical issues that it raises, and that can come into play in affidavits because we’re going to have a situation where a lot of exhibits have been created by AI, a lot of the appendices and things and annexures, and that was one of the things that we also raised with the making sure that the Practice Note was going to work. And then the third thing that we said in our article is, “well, we actually think we have a responsibility to teach students something about how to use AI because they’re probably going to be expected to use it in their workplaces, if not now, then very soon, and they need to be actually able to do it well.” And I think this is where it is very complicated because we don’t have the expertise yet to understand how to use AI well, I don’t think as a society, I don’t think any law school can really claim it. We’re all working furiously at it. And I think it’s a very American dominated technology. A lot of the training has come out of American databases, and so a lot of AI have been trained predominantly on American materials, and so they tend to have this kind of, what I think of as this somewhat overconfident kind of tone that – and I think that’s actually probably one of the biggest problems with AI is that – when you or I decide how we’re going to express something, the nuance that we have in terms of the level of certainty, it’s like best recollection versus verbatim versus gist; how certain am I? That comes through in my tone. So I might say, “well, I think,: you know, like I’m talking to you about AI and I say, “well, I think this is what’s going to happen.” And so you can see my level of uncertainty. Because AI is actually completely brainless, – well, generative AI, large language models – has no idea. It’s picking its level of certainty based on either what it thinks you’ve asked for, or what most people have said in the training databases that it’s getting its material from, and that’s actually in no way reflective of what your actual level of certainty is for what you’re saying. And so that message isn’t necessarily coming through with anything like the right nuance. And so I think one of the things that we need to learn to do as lawyers is not just to check for content, like, “we didn’t make up any cases, did we?” Perhaps with fake citations, which people are really starting to get in trouble for in the courts because AI does do that. But also, that the tone that we’re delivering every element of our message is right. Because like you say, it’s a persuasive business law. So, if I’m giving a bit of evidence, or if I’m talking about something, or if I’m making a submission on something, are my strongest points the ones that I’m the most confident of coming across with that level of nuance? Or am I putting forward a weak point terribly confidently because that’s what ChatGPT tends to do, when really I don’t have that level of confidence, and to put that argument with that level of confidence is actually potentially misleading of the court? So I think that one of the things we need to learn as lawyers using a generative AI is that it’s not just enough to check the cases haven’t been miscited. We really have to, like you say, understand our case and think about the level of certainty that we’re, or the nuance, the persuasiveness with which we’re delivering our message. |
| 00:57:02 | DT: | That’s such a good point, and I think one that’s easily overlooked. We’re at a stage in the adoption of this technology where we’re overly concerned with the accuracy of the content, right? Because it can be wrong. We’re like, “oh, well, that’s not fabricated. Great.” |
| 00:57:16 | SW: | “That case really exists. Brilliant.” |
| 00:57:17 | DT: | “Tick, I’ll add that.” But you’re right. The medium is the message, right? The way we present the message is part of the message itself, and I think it is easy to have that suggested to you in that form. “That is how I meant to express that, right?” |
| 00:57:32 | SW: | “Sounds good. Yeah.” People read things and they go, “oh, that sounds really good, sounds really persuasive. I was persuaded.” And you think, “yeah, but should you have been? And were you appropriately persuaded about the right bits of this?” |
| 00:57:43 | DT: | Yeah, that’s right. I hope we all think about that. Yeah. |
| 00:57:50 | SW: | Yeah, no. And again, we come back to that point you made so well earlier on, where we’ve got limited time and so, in an ideal world, and you and I were lucky, we worked at big firms which had big matters worth lots of money and big budgets, but you know, what happens when you are working in very much cost and time constrained practice, or you’re a self represented litigant and you plug it into ChatGPT and it sounds really plausible and you’ve got no idea because you’re not a lawyer. Like how badly should you be punished for the fact that you’ve just delivered a whole lot of garbage to the court? I mean, it’s possibly no worse than the garbage you would have delivered. |
| 00:58:24 | DT: | Yeah, I think that there is an argument. Is there a harm reduction element there better from chat GPT with the occasional fabricated case than from a Facebook group with – |
| 00:58:35 | SW: | Conspiracy theorists and pseudo law. |
| 00:58:37 | DT: | Based entirely on the Magna Carta, right? So – we could have a separate hour long conversation about that. |
| 00:58:42 | SW: | I love pseudo law. It’s such a fun topic. |
| 00:58:44 | DT: | Yeah. It is. And yet at the same time, I don’t know, it’s one of those things that remind us how unattainable legal information can be for so many people who experience a legal issue and that going to the Facebook group that signs names in blood with triple brackets around them so that the tax office can’t collect tax from them. |
| 00:59:05 | SW: | It doesn’t work by the way. |
| 00:59:06 | DT: | It doesn’t work. We should clarify that. |
| 00:59:08 | SW: | I was putting on my tax hat for the moment. |
| 00:59:10 | DT: | Yeah. You do have that hat. |
| 00:59:11 | SW: | I’m going to say, no, it doesn’t work. |
| 00:59:13 | DT: | But the reason someone might go to that group, which we’re all happy to mock, is because the alternative, which is high quality, correct legal advice, is unattainable. |
| 00:59:23 | SW: | Yeah. |
| 00:59:23 | DT: | Sonya, we’re nearly out of time. Before I let you go, as always, I want to finish with a question for our listeners who have recently completed their law degree or are perhaps still studying it. On the basis that affidavit drafting isn’t entirely handed over to large language models in the next 10 years, for those who are starting to develop their skills in drafting an affidavit, and I suppose we can expand that out to persuasive legal drafting more generally, following on from the tail end of our conversation, what’s one tip you’d want to leave them with? |
| 00:59:51 | SW: | Well, I think it’s a difficult area and so I think don’t underestimate it, but I think the best thing for lawyers to remember is that their paramount duty is to the court. So, your first job is to assist the court. So, when you’re drafting anything, including an affidavit for the court, you have to think about what the judge deciding your matter needs to know and what issues they have to decide and how you can succinctly but comprehensively communicate the relevant facts and the law to the judge so that you can help them decide, preferably in your favour. And then I think it’s really important whenever you’re drafting anything, to think about, who is your audience? What do they need to know? What is this document doing for them? And I think you can apply that to affidavits, you can apply it to pleadings, you can apply it to anything that you might draft, and not just court documents either, like I think that’s something for drafting for all lawyers, that it’s really easy to forget who our audience is. So I think, yeah, who are your audience? What do they know? What don’t they know? What language will they understand? What language won’t they understand? Are they legal experts? Are they not? And then, of course, with an affidavit, there’s that extra issue that you’re not actually drafting it on your own behalf, you’re drafting it on behalf of someone else, and it’s really their statement. And so that puts another layer on that, you have to be empathic, step into their shoes, be them when you write, but a really superior, legally trained version of them. |
| 01:01:14 | DT: | Well, and I really like that. And I’m going to count it as one point because it’s really two sides of the same coin, isn’t it? Paramount duty to the court and, know your audience, is really the professional responsibility side and the persuasiveness side of the same point, which is who is going to read this and why. And I think keeping that in mind is going to help us avoid that greatest of sins that we’ve talked about today, which is drafting to satisfy what you need to satisfy to win rather than drafting to give a true account of the conversation. Sonya, thank you so much for joining me today on Hearsay. |
| 01:01:45 | SW: | Thank you, David, for having me. It was a pleasure. |
| 01:01:57 | DT: | As always, you’ve been listening to Hearsay the Legal Podcast. I’d like to thank my guest today, Sonya Willis, for coming on the show. Now, if you’re looking for some more litigation or court process related content, go and check out our episode with data analyst Hans Weemaes about causation and how you can make better causal inferences, or question the causal inferences of your opponent. That one’s episode 129 called ‘Grasping Causation: A Data Science Explanation of Causal Inference and the Role of Counterfactuals.’ You could also try another useful episode in the litigation context, that’s episode 127 with Caroline Hutchinson called ‘Legal Liaisons: Navigating Court Communication Protocols.’ If you’re an Australian legal practitioner, you can claim one continuing professional development point for listening to this episode. As you know, whether an activity entitles you to claim a CPD unit is self assessed, but we suggest this episode entitles you to claim either a professional skills point or an ethics and professional responsibility point. Take your pick. You could also claim it as a barrister’s skills point if you happen to be a barrister. If you’d like more information about claiming and tracking your points on Hearsay, head on over to our website. Hearsay The Legal Podcast is brought to you by Lext, a legal technology company that makes the law easier to access and easier to practice, including your CPD. Finally, I’d like to ask you a favour. If you like Hearsay the Legal Podcast, please leave us a Google review. It helps other listeners to find us and that keeps us in business. Thanks for listening and I’ll see you on the next episode of Hearsay. |
You must be a subscriber to access this content.