Join the NALP team as they interview Robin Ghurbhurun and discuss the impact of AI and how it could impact the work that paralegals do on a day to day basis.
If you want to register for upcoming webinars and events, visit our Events page to see what is happening and when.
Transcript
0:04
So I’d like to say welcome to this NALP webinar looking at AI and its impact on the legal sector and in particularly how it’s going to impact paralegals. I’d like to introduce the team. So my name is Chantelle Cook. I’m one of the directors of NALP. Also with us is Amanda Hamilton, our former CEO and patron of NALP, also for some reason labelled Jane Robson. But I can assure you that it is Nigel who looks after our marketing at now he’s here too. Nigel, kind of be in the background, if you’ve got any
0:35
problems, feel free to direct message him on the chat. So we’ll be using just the chat, not the Q&A or anything like that, just the chat section. So if you got any problems, message Nigel on the chat, what I’m trying to say and also if you’ve got any questions during the webinar today, don’t let them kind of distract you. When your head, pop them into the chat and we will come to them a little bit later on. We’ll be trying to answer all of the questions. So I say if you’ve got a question that’s burning, just pop it in the chat and we will come to it in a little bit and Nigel will be dealing with the questions as we go through. And plus we also have of course our experts today, Robin Gerber and who I will introduce properly in just a few minutes. But
1:16
first of all, a few housekeeping things. As I mentioned, videos on or off, Your choice, microphones off, please. The webinar is being recorded. Just so that you’re aware,
1:28
we’ve got a number of other webinars coming up in the future and I will run through those. But just so that you know immediately for your Diaries while you’re here, the next one is actually next Wednesday and that’s going to be looking at if you’re interested in a career in probate research, then this is the webinar for you. So we’ll be talking about that next Wednesday, but I’ll give you some more information about the webinars that we’ve got booked up right through until the summer, so you can get those into your Diaries a little bit later on. OK. I think it’s time to start.
1:58
This webinar is all about providing insight on AIS current and future impact on the legal profession. Plus we’re going to have some advice on how paralegals can open their own roles by working alongside AI. Our expert today is Robin Gerbrand. He’s the Managing Director for Further Education and Skills at the UK’s Digital Data and Technology Agency for Tertiary Education, Research and Innovation. And that’s a bit of a mouthful, but it has a nice short day disc. So I’d like to
2:28
welcome Robin. Robin along. Thank you very much for joining us for this webinar. Robin, I know you wanted to give it a little context before we get started, so let me pass over to you for a moment. Thank you very much, Chantal, and thanks for the kind introduction. Delighted to be here and to lead this session on naps webinar on AR. I must confess though, I will start by saying I’m not an A I expert. Although my organisation Jisc does provide leading insights and support for the effective and responsible use of AI across tertiary education,
2:59
which is further and higher education and our research institutes critically, I think it’s important to to just provide some clarity on AI itself. I think for any individual contemplating using AI, they must develop a solid understanding of of data literacy as a foundational skill for AI. And AI should not be viewed as a standalone skill which is just another tool. So it should be part of a broader set of competencies that include communication,
3:30
collaboration, critical thinking, data literacy, digital literacy and problem solving skills. I’m sure all of those that that many of our audience have here today. Gen AI or generative AI is a subset of artificial intelligence technologies and it’s capable of generating new content that at times resembles human like outputs. These systems often leverage vast datasets to produce text, images, code, and more.
4:00
And they exhibit an understanding or mimicry of human like creativity and reasoning. Unlike traditional AI that primarily focuses on interpreting or classifying existing information,
4:13
Jen AI takes more of a proactive role in tackling tasks. It can generate new content based on learned patterns, summarise large documents, and interpret chart trends from unstructured data as the like. But AI does do that with limited guidance. So my thoughts today will be shaped by several sources including Gartner, my own legal team at JISC and Workflow GPT that recently published a
4:43
the future of Gen AI and legal work. So Chantelle, back to you. Look forward to the questions. Thank you. I guess, although you’ve kind of given us a context for a I, is there a definition that we should be aware of?
4:56
I think that it’s anything that is produced from machine learning or all of this computer based generated information. So anything that’s not human generated I think is the simplest way to look at artificial intelligence. It’s not human intelligence. It very much is generated through computing.
5:19
Thank you. So how is AI already in impacting the legal sector? In other words, where can we see it at work already? Who’s using it? How are they using it? Can you give us some insight into that? Yeah, great question. I mean, clearly it’s a rapidly growing market and there’s a loss of newly emerging software and you tend to get fight or flight when people come across anything that’s new and emerging. But we also have well established software providers that are looking to add value to their current offerings. So if you think of Adobe and Adobe PDF
5:49
Memory for example, what it does, it summarises through its own AI assistant and allows you to engage with a document by asking it questions with Instant answers linked to sources primarily from within your own document
6:06
and can be used to craft emails, presentations and more. It’s also got a generative summary function that pulls up the key points. And then we have the likes of Co Counsel Core which is from Thompson Royces. Now this is a Gen AI legal assistance and it was launched in the US last year and it equips customers with kind of eight Gen AI powered core legal skills and that includes things like preparing for a deposition, drafting correspondence,
6:36
searching a database, reviewing documents, summarising a document and extracting contract data and contract policy compliance and timeline. Now we’ve seen law firms such as Clifford Chance, they’ve they’ve announced pilots and trials in collaboration with major software companies so for example with Microsoft and using Microsoft Viva and Copilot. And these mostly remain internal to law firms work rather than involving client service
7:06
delivery or indeed kinds information. At present, there could be a danger of fatigue. There were so many software companies that are claiming to they’re about to revolutionise the legal sector
7:20
and the most common applications that we’re seeing are around risk identification and prediction, administration, profiling and text generation. Now if I cite my own organisation, our own legal team at Just, they’ve recently implemented an AI solution that provides that’s been provided through a law firm and it’s developed in collaboration with a major AI platform. And they’re using this to accelerate contract management reviews and also exploring applications in contract management.
7:51
So we know the legal sector is already starting to explore and find opportunities to leverage it for efficiency, for accuracy and for innovation. And if you look at research from McKinsey in 2022, it states that 50% of all organisations in surveys were reporting using some form of Gen AI. So legal teams and firms really need to think about ways in which they can align themselves with the technological upheaval that we’re seeing happening across many industries.
8:22
And at the end of 2022 it was reported that 3/4 of the larger solicitors firms are already using AI’s, nearly twice the number from only three years ago.
8:32
60% of large law firms are at least exploring the potential of new generative systems and 1/3 of small firms. So research firms are sketching a future where Gen AI not only augments the capabilities of legal professionals but it introduces new paradigms and equal practise. And I think this begins with conversation AI and then extends to an entirely new AI generated applications. And the potential transformation is vast. I think the rapid pace of evolution
9:03
does require caution, and in any of these areas it comes with a health check, especially when choosing foundational AI technologies to kick off your own organisations, AI journey and law firms that are poised to thrive in that generation Gen AI space, they kind of choose one of three pathways. Do you buy one off the shelf piece of software? Do you privately label your own GPT or do you build your own private GPT from scratch? And I think whatever
9:34
path you choose, you need to consider the present as well as the future in terms of that desired outcome. Back to you, Chantelle.
9:43
I guess I’m a little bit surprised. It’s being used so much already and it’s so prevalent already. And that then has to say, so how’s that gonna change in five to 10 years? Because if we’re already here and it’s that prevalent, what’s it going to be like in the next decade? And in particular, how’s that going to impact jobs? Indeed. Well, I’m not a futurologist either. It’s a bit like trying to trying to determine the weather at the moment. I think it’s fair to say if data is the new a new oil, then AI is the new electricity. In other words, it
10:14
finish in the same way that electricity revolutionised industries by providing a new energy source. So AI is definitely transforming how we approach, execute and innovate it in in our work. So the capacity to automate up to 40% of the average work day, what are the implications for efficiency and productivity? They’re quite profound and certainly in sectors that are reliant on the processing and analysis of vast information volumes such as the legal field. Now I think the legal profession stands at that pivotal point or moment,
10:44
the time where Gen AI is poised to to transform the landscape of the legal work. Fundamentally. So law firms and legal professionals are encouraged to embrace the shift, not fear it and not see it as something as in in a distant future, but as an unfolding and reality that that demands engagement, exploration and
11:06
temptation. And you have to think about taking a risk based approach to anything that’s innovative and thinking about what is the risk appetite within your own organisation. And let’s face it, the legal profession is almost centred on on risk and risk appetite. So the greatest impact will be speeding up response times across the industry. So especially when summarising large swathes of information, full adoption of dynamic contract management and automation, capable of updating business intelligence and responding to those changing risk factors.
11:37
I think if you look at the next three years, let’s say something, let’s go out to 2027. According to Gartner, I think by 2025, next year 30% of enterprises would have implemented AI augmented strategies, which suggests a move towards a more sophisticated legal kind of analytics and predictive modelling and enhancing precision in case outcomes predictions and legal risk assessments. The year after 2026, the integration of AI colleagues I know from the education sector, we already have universities and indeed colleges having AI
12:08
instance almost Alexa in the room supporting them and working with the teachers and professors in terms of supporting the learners. So the idea that you’re gonna have AI assistance is not so far fetched and legal research, case preparation and even mundane administrative tasks such as enabling lawyers to focus on higher value activities. Further out to 2027, the emergence of applications that automatically generate and generated by AI without any human involvement that will revolutionise legal software development,
12:39
make custom solutions more accessible and probably more affordable for law firms of all sizes. Now thinking about a paralegal profession would be remiss of not, you know, focusing in on a paralegals. I think they’re involved. Involvement should be more nuanced and kind of nuanced business LED advice rather than collation and administration of data due to deployment of the AI tools. So we’ll focus on quality control and the optimization and safeguarding of contractual arrangements. And there’s also a great opportunity
13:11
to upskill paralegals. So if you think about prompt writing, the opportunity now to support paralegals in the effective training and deployment of legal AI. So if you think about the legal sector being traditionally rooted and precedents and and accurate documentation and it’s experiencing major change in that paradigm with integration of Gen AI. So crucially, front engineering there is more. This is more than a trend. It’s a fundamental shift which you’ll see
13:41
reshaping the way in which legal professionals interact with technology to enhance their work. So in law the stakes are high, you know a poorly constructed prompt could lead to misinterpretation or or legal and accuracy, whereas a well designed prompt can provide accurate, reliable results that are critical to legal decision making. So consider for example, an AI tool used for contract analysis. A well designed prompt can help the tool to not only identify key clauses
14:11
and terms, but also understand their implications in different legal contexts. The capability transforms the tasks such as contract review and due diligence and even legal research, saving countless hours while improving accuracy. So for a paralegal, complexity and skills are required. Legal prompt engineering is multidisciplinary. It requires a deep understanding of both the law and Anna and and AI. What a great opportunity for paralegals to be at that pivotal role
14:42
and and understanding both AI and and and the law. So mastering the film requires not only familiarity with legal terminology and concepts, but also understanding how AI models process and respond to language. And this dual expertise can be challenging, but it’s essential to develop the actual development of you know, effective legal AI tools. Worth noting though, like with anything you know, when it comes to computing and processing, Gen AI can only produce new content
15:12
by recycling what’s already and what it’s already been trained on. So garbage in, you’ll get garbage out, and the quality of output is often determined by the quality of the input in the 1st place. Back to you, Chantelle.
15:24
That’s very interesting because when you were first talking, I was kind of thinking about all these sort of often more administrative tasks and things that that paralegals within solicitors offices in particular do. Sounds like a lot of those might go, but actually from what you’re saying, yeah, they might. But in order for them to go, you need somebody who really understands what’s going on in order to input that, which means you might lose a job here, but you’re gaining a job there. And actually that job could really only be done by somebody
15:55
who has got the expertise for paralegal. Absolutely. And that’s the whole point about augmenting the professional. You know, you’re not, you’re not looking at displacement, You’re looking at how do I augment my skill sets and how do I further develop myself as a professional in this film. And that’s not just the legal profession. It’s a number of sectors that that we’re seeing at the moment. And based on that, then what advice would you have? Because I love the idea of understand AI and augment my skills.
16:20
How? Yeah, OK, so I mean again it’s emerging but in most systems. So we think about most systems that already employ AI, the technology there acts as a as a support to improve the work of humans. So where firms are using such tag, there are signs that increasingly familiarity is not only helping them to use it effectively. So playing with it, trying it in a safe environment, but it’s also overcoming those concerns that the use of AI, as you said, could some way replace the humans
16:50
you know and and and and you forget about the number of jobs and different types of jobs that might be created across a number of professions. So most surveyed workers say that AI has improved both their performance and their working conditions. So if the AI, if the paralegal is the human interface with the AI, then that’s going to be essential for the foreseeable future, particularly in areas such as, you know, there’s something called AI hallucination and I’ll give you an example. So identifying the AI
17:22
hallucination in terms of the detail and responses, and verifying those outputs that have been generated will be a critical human trait. So, for example, ChatGPT can be prone to hallucinations or inaccuracies. In one example, ChatGPT falsely accused an American law professor of sexual harassment and cited A nonexistent Washington Post report in the process. Now in February, a proposition A promotional video from Google, and they’re right. And Google’s rival to chat,
17:52
PT, was barred. It’s called barred BID. It gave an inaccurate answer to a query about the James Webb Space Telescope, and it raised concerns that the search company had been too hasty and launching A riposte to the launch of ChatGPT. In other words, open AI’s breakthrough to Chat bots are trained on a vast trove of data taken from the Internet, although those sources are not available in many cases. You know what we’ve seen in the education sector where students have produced pros and often
18:22
the process includes references and citations and actually the fact that that it was almost as if it still had the sticker label on it from from someone else that had written it and the student was claiming it as their own. So you can imagine what that means in the legal profession. So the fact that they’re trained on on, on all of this data,
18:44
it operates like a predictive textile. So how many times have you text someone and then it starts, it’s changed it to what it thinks it is, what you want to say, how annoying it becomes unless you turn off that predictive tool. So it builds a model to predict the likeliest word or sentence that come from the user’s prompt. And this means factual errors are possible, but the human seeming response can sometimes convince users that the answer is correct. So paralegals I think could verify the identity of clients, catch fraudulent transactions,
19:16
you get AI phishing scams you know voice phishing scams. Again the human side of that is critical to to doing that quality assurance and also defining where the legal liability lies in in a in the AI value chain. So you you will have you know a number of of of of contributors, creators and deployers and and if you don’t have the human checking whether for example in the value chain that supported the final document
19:47
GDPR
19:50
intellectual property in A whose checking on all of that in terms of what the AI is generating. Because a I can’t make those distinctions itself, but a paralegal can.
20:01
So it sounds like which I’m very pleased to hear, there are plenty of opportunities out there for paralegals despite the impact of AI which obviously is going to change things completely. I guess another little thing that you just kind of vaguely touched on there is shouldn’t be using a I to do your qualifications, should you? No indeed I don’t know why. I think, I think as with any sourcing and referencing and and and acquisition of information and knowledge, we all use sources whether that’s on the on the Internet,
20:31
ohh, that’s Wikipedia, whether that’s my book, whether that’s from a colleague. I think when using it, it’s then how it’s citated and how it’s referenced. We’re more importantly how it’s applied to the context in hand. And I think you know that’s where in the legal profession for example paralegals looks at how they can target AI rather than some of these novelty factors. We’re all playing with it, me included, various points, and then you hone in on actually where are the greatest benefits. And I think it could be really powerful from an education and assessment perspective,
21:01
but it has to be measured, it has to be balanced and it has to be ethical as well, of course.
21:06
Absolutely. Thank you very much. We’ve got a little bit of time left and I’d really love to hear your questions. So if you’ve got any questions, please do pop them in the chat and Robin will be hopefully be able to answer them here. I know he’ll certainly do his best. This is such an interesting area. And obviously, as Robin said, he doesn’t have a crystal ball any more than any of any of us do. But it’s definitely an area that it’s worth understanding, worth getting your kind of head around as much as you can. And then because there are all these opportunities and because if you’re not,
21:37
it’s like a lot of things, if you’re not kind of getting your head around it and you’re not at least trying to embrace it, then you could find yourself a bit left behind at the end of it. So definitely, please put your questions in the chat and we will put those to Robin in just a second. I suppose the other thing to ask you, Robin, is
21:55
do you,
21:56
I mean, do you think that although AI is moving really, really fast and you had your kind of three-year little plan that you talked about and not planned, but you know the kind of predictions for the next three years which seem I think three years out seems reasonable? It seems reasonable. But do you think there’s any way in which we can understand what it might be like in 10 years? Or do you think we’re going to see a massive slowdown and actually it won’t be quite as radical as maybe it all feels like it will be? No, I think it’s fair to say that it certainly won’t slow down. I think
22:26
they want it’s familiar with Moores law and technology and semiconductor technology and the pace of change of computing generally because of the massive information that is currently available that can be ingested. And you see the investments that are being made in the storage, the cloud solutions even in in in in the machines that are required to process this vast amounts of data. It’s being backed heavily by capital investment in terms of them producing
22:57
the number of workflows that are required for a number of industries whether that’s automation on the shop floor and and digital twins through to a textual information. I think going back to your question about how paralegals can best prepare themselves for the future. I think that’s the gift that the those that have got an interest in in the legal and paralegal profession can absolutely wrap around. And and I think if I was a paralegal today I’d be looking to understand how automation
23:27
to benefit the firm I worked for as as part of an in house team have have an insight on the day-to-day processes that are required to keep things running effectively.
23:37
Things like exploring opportunities for paralegal to be in their vanguard of AI experts. So understanding where AI adoption will bring the most benefits and who have an understanding of the limitations and also those pitfalls. And sharing where it works and doesn’t work and pointing to what does good look like and and best practise although best practise yet to be determined. You know, each one person’s best practise could be tested by another And remembering that the impacts of impact of AI is going to be across all industries, you know we’re not alone here,
24:08
it’s it’s impacting. If you look at the World Economic Forum and I highly recommend looking at the World Economic Forum publications on future skills, they look at the demands of the skills that are as citizens across the globe that are going to be needed. And some of those skills are the skills that inherently make us human,
24:25
you know, and thank God you know. I always use the example of a of the AI and a medical profession where you look at the diagnosis and even perhaps the proposed solution or or healthcare remedy won’t replace for the foreseeable future the empathy and the bedside manner of a GP or consultant. When they’re talking to the patient. What it may do is hello, accelerate getting to the diagnosis. So those skills that make us more human are the skills that very much employers are looking for. And if I was in any industry, I’d say
24:55
develop those human skills. And finally, don’t think you can fill in the gaps
25:02
that the AI isn’t currently doing. Because eventually what happens when the AI fills those gaps, You know, if you focus in on, we’ll just do the bits I doesn’t currently do, actually do the things that, you know, make us, you know,
25:16
cognitive social constructs as human beings. I’ll stop there.
25:21
Probably we’ve got some questions coming in and please we’ve got a few more minutes. If you have any other questions, pop them in the chat. I think Nigel is handling the questions. So I’ll, I’ll hand over to you.
25:30
And yes, we’ve got two questions from Tanya. And first one is what specific AI technology is being used in in the sort of the paralegal world. And so I think she’s after a couple of things to for a couple of recommendations for things to look at. And then also what are the best practises for training paralegals to use AI effectively?
25:54
Yeah, interesting. So I mentioned one or two that that like Clifford Chance are using with Microsoft and and obviously using Copilot
26:03
and some of the Adobe software. I think they are going to be at the core and the mainstream and and you know the software solutions are developing outside of those companies. They’ve got huge amounts of investment that they’re making internally probably won’t be made available.
26:20
Study lessons are made available through their supply chain. I think when you look at some of the solutions, there will be legal automation tools. So they’re kind of streamlining repetitive tasks such as document filing and case management. There’s legal research platforms. These platforms will provide lawyers with relevant case laws, precedents, legal theories I guess around unprecedented speed,
26:44
large language models. I think what we’re gonna see is smaller language models actually specialists at the moment chat GP GP’s are large language model for example. But I think we’ll see start to see smaller bespoke language models based on demand from the consumer and they will assist with things like drafting legal documents, contracts and briefs, predictive AI. So we’ll look at historical data and then those tools will look at potentially the potential case outcome
27:11
and it will help legal teams to strategise formulation. And then there’s contract review systems and there are, I mean there are, there are many and I’m I’m not going to make a recommendation because I’m not a legal profession, but I think it’s worth exploring with your legal peers as to who’s using what either through those kind of communities of practise.
27:31
Hopefully I’ve given a sense of those main tools, automation tools, research platforms, language models, predictive AI and those kind of contract review systems type, any of those into Google or your preferred search engine and I’m sure you’ll come up with some some software applications that are being created already.
27:51
Thank you. I think we’ve got, we’ve got a couple of questions coming which are which I think maybe the first one you’ve kind of covered, but I’ll mention it anyway. Are there any websites or resources you would recommend as an introduction to how to train AI and understand how it works under the hood?
28:07
Umm,
28:09
yeah, actually there are some free resources which I always highly recommend,
28:17
and the ones that have just recently actually in the last week or so that have been released. If you’re new to the world of AI,
28:28
there’s about 8 to 10 resources. So Google offers AI courses and they’ve got five different courses and you can learn about Jen AI and there’s an introduction to AI and it leads to a kind of solid understanding. So do look at the Google AI courses. Microsoft have got on a course that covers the basics and then more and it starts off with an introduction and continues through learning. Now I can provide you post this meeting and if it’s alright with help, I’ll give you. I’ll provide links to all of these. You can just click on them and I’ll take straight to these.
29:00
These these resources prompt engineering that I mentioned with ChatGPT. It’s a six module course and that’s actually created by the Vanderbilt University and offers beginners a starting point to write better prompts. And it starts by learning effective prompting, and then complete the course knowing how to bend, to bend ChatGPT to your will, dare I say, you know, using it in that way.
29:23
There’s also open AI that created ChatGPT. They’ve gone into collaboration with deep learning, which is Google kind of deep. Google minds offshoot and they’re offers offering a course that’s taught again, all three. And it starts off with best practise and it finishes with a better understanding of prompt hands on practise. So there they are, two or three ones. There’s one called Big Data, Artificial Intelligence and Ethics and that’s the University of California. These are coming from the states again
29:53
completely free and it and it introduced you to IBM’s Watson which is kind of the initial nuclei I guess of of of chat bots and again it starts you off by learning about big data opportunities and knowing the end limitations with AI. So there’s there are some wonderful courses out there that I’d highly recommend. So I’ll drop the 8:00 to 10:00 or so links to those those courses and you can export your own will and and the beauty, as I said, they are free to help your, your own understanding.
30:24
That’s brilliant. Thank you. And that kind of answers the next question which was about do you think cost is going to be kind of a strict entry, But actually if there’s so much free training out there, I think that the answer is no, it won’t do that. We are actually running out of time. I’m really sorry. But thank you so much for all your advice and information that you shared with us about a I. It’s a massive area and I’m absolutely sure we’ll be doing this webinar again maybe in a year’s time and it will be completely different. Then there’ll be lots more,
30:54
lots more to learn. I know we’ve got a couple more questions and I’m sorry that we just don’t have time for those. But there will be a recording of the webinar available on the website for now members. And then later on that becomes available to non members as well. But that will be on the website in the next sort of week, well, within a week certainly. So you can always listen to it again and hopefully fill in anything else that you maybe just didn’t hear the first time round. Thank you so much, Robin, and also for your offer to send over those links. I’m think they would be fantastic
31:25
for us to share. So I think I very much appreciate that. So thank you so much for giving us your time. Just very quickly, I know we are running out of time, but just a quick announcement of the events coming up next week. We’ve got Danny Curran, who’s the CEO of Finders International. He’ll be answering questions about having a career in probate research. On Wednesday, the 5th of June, we’ll be talking with Hillary Briggs about how to overcome overwhelm. So sometimes you feel a bit overwhelmed by how much you have to do. Hillary has a way to get past that and get
31:55
everything done that you do need to get done. And also on the 11th of July, that’s the Thursday, we will be speaking to Barbara about how to convert your paralegal experience and qualifications, if you fancy becoming a solicitor. So those are the webinars that we’ve got lined up already. You can find all of those on our website, nationalparalegals.co.uk/events or navigate through it through the About tab as well. That will also get you there. So thank you so much to everyone for joining us today. Thank you very much
32:26
Robin Berberian from Jess for sharing his information about AI and we wish you all a lovely afternoon and we’ll hopefully see you next week.
Previous webinars can be seen on our YouTube channel here.
