Intro: 0:01

Welcome to the Actionable Futurist podcast, a show all about the near-term future, with practical and actionable advice from a range of global experts to help you stay ahead of the curve. Every episode answers the question what's the future on, with voices and opinions that need to be heard. Your host is international keynote speaker and Actionable Futurist, andrew Grill.

Andrew Grill: 0:30

Not surprisingly, the latest Gardner hype cycle for AI puts generative AI at peak hype. Even my parents in Australia are talking about it. So how do you cut through all this chat GPT hype, particularly when it comes to the context of professional services? To address these issues and more, my guest is Jaeger Glucina, the Chief of Staff at World Leading AI Company Luminance Welcome, Jaeger.

Jaeger Glucina: 0:51

Hello, thank you very much for having me, Andrew.

Andrew Grill: 0:53

For those of us that haven't heard about luminance what does luminance do and what problem do you solve?

Jaeger Glucina: 0:58

Luminance is an AI which really revolutionizes how lawyers work and I know that is a word that is majorly overused revolutionized but it quite literally reads a whole contract, works out the risky parts, which bits a lawyer needs to think about, and can even read draft a contract with a simple prompt from you.

Jaeger Glucina: 1:17

It combines generative with analytical AI to do something very difficult that no one else has really managed to achieve, because actually AI is an incredibly hard thing to do. What gives us a major advantage is the fact that it's been trained and validated by the top law firms in the world. So this is AI that performs specialist legal analysis and that is a radical change in how lawyers are going about their work today. If you think about it, there are people in every office and every building in the world dealing with contracts every day whose lives would be dramatically improved if they had this technology. We have customers in 60 countries using Luminance to augment and automate big parts of their daily work, like drafting and negotiating contracts, which is really key to alleviating the burden that lawyers are shouldering. So Koch Industries, lamborghini, yokogawa, Hitachi and Liberty Mutual are just a few of the businesses using us right now.

Andrew Grill: 2:09

What's the difference between generalist AI platforms such as ChatGPT and specialist AI platforms such as yours, and what are the benefits of each? Don't get me wrong.

Jaeger Glucina: 2:18

Chatgpt is an incredible development in the field of AI. It's brought AI into the public consciousness, I think for the first time properly in many, many years, and it can do really interesting things. So many interesting things, but two very basic examples it can write a jingle for an ad or even write a pretty decent sounding poem. I know we've all experimented with that, but it's generally and I emphasize the word generally here quite useful at doing very generalist things. So I like to describe it as that person we've all sat next to at a dinner party who seems to know something about everything and they're totally convincing. They've got the whole table hanging on their every word. But afterwards, when you leave the party, you think to yourself hmm, I actually think that might have been a load of garbage.

Jaeger Glucina: 3:02

So that's essentially what ChatGPT is aiming to do. It's aiming to give an answer at all costs, without a thought for accuracy. Numerous is quite different to that. The reason being is that over the past few years our team has been at the forefront of specialist legal large language models or LLMs as we all know them now and we've been working with some of the world's top law firms and in-house legal teams to get a really big head start and we've been able to, as a result, develop an AI which is an actual legal expert.

Andrew Grill: 3:31

I have an alternate view of how to describe ChatGPT. I liken it to an enthusiastic, always on intern and, just like your dinner party guests, they would go off and research things. They would be very enthusiastic, they'd come back with some of the answer but, as you know, you'd always have to check their work. You would never give an intern's work to a client directly or publish it online. So, dinner party guest who knows a lot about everything, or enthusiastic intern, I wonder if they were both at the same dinner party what answers would get?

Jaeger Glucina: 4:00

That's the question.

Andrew Grill: 4:01

So you started your career as a barrister and solicitor in New Zealand representing earthquake victims against insurance companies. So could you explain your journey from New Zealand to the world of AI in Cambridge?

Jaeger Glucina: 4:12

I started my professional career in what I personally found was a very rewarding field of law. Others might call it boring, but I'm very grateful for the experience that gave me and the fact that I was able to do some good for my clients. A lot of my time, as you can probably imagine, during that role was entrenched in discovery work, which is essentially reviewing every document, email, report, text, anything really that could be relevant to a court case, and we're talking the days of physical filing and printing of literally everything. Most of that's now moved to digital, but I experienced firsthand some of the real challenges that still face lawyers today. It's the sheer volume. So we all know that data is growing every day. We're producing more information and content than ever and it's only going to keep getting bigger. But you know, what's not changing is the deadlines set by clients and by the courts. If anything, they're getting shorter. So burnout is a real thing in the legal profession.

Jaeger Glucina: 5:05

But my experiences early in law highlighted how the quality of my work and my work-life balance could be improved by technology which could automate legal tasks. Tech like that didn't really exist at the time, or they're in their very early years and hadn't made their way to New Zealand? That's probably more likely. So when I moved to London shortly afterwards, I spent a year in-house at another tech startup, where much of my day-to-day job consisted of generating and reviewing contracts and submitting companies house filings for our customers. Like many startups, that company didn't make it, but I'd been looking into legal tech and had come across something that sounded very exciting. It was a new AI product which was set to change the game for M&A legal due diligence. I'd been burned before by startups, obviously, but I did some pretty thorough research and it quickly became quite clear to me that this was really something that had the potential to change the future of the profession. That was Luminance, of course, and that's where my introduction to AI started six years ago.

Andrew Grill: 6:04

You said it's skilled, specialist, legal grade. Ai is the future of business, but what does this mean in practice?

Jaeger Glucina: 6:10

Well, luminance employs a large language model LLM, as we've established. The key difference from other LLMs is that Luminance has been purpose-built from inception for legal, specific applications. We talked a little bit about that. We were founded in 2015, so since then, the AI has had the chance to have a few years to be informed by over 150 million verified legal documents, which is very different to any old document that's been scraped from the internet. It's that exposure to the vast numbers of specific legal documents, combined with our core proprietary AI deep domain knowledge and constant use and classification by the top legal users all over the world. That means it is now that most advanced legal LLM and therefore the most reliable, which is the important part.

Jaeger Glucina: 6:57

That's where the limitations of DPT and other generalist AI chatbots begin. They're not fit for purpose in very specialist fields. Take medicine, manufacturing, engineering. All of these require that same level of deep domain knowledge that legal does. It's important that the AI being used in those fields is able to adapt to and learn the nuanced ways in which those industries do business and make decisions. Anything less than that has the potential to do more harm than good.

Andrew Grill: 7:24

So I do a lot of work with legal clients and they all want to know about AI. I've spent many an hour or two talking to law firms and their clients. They're all a bit worried about AI because if you charge by the hour and AI can do things in milliseconds, this will disrupt things. So how do you think AI will disrupt the role of a lawyer both now and in five years time?

Jaeger Glucina: 7:45

Great question. A lot of lawyers have and this is in my experience been historically quite worried about and to an extent, some still are the AI overlord coming to take over some time in the near future and, among other very scary things, putting everyone out of a job. My response to that is that we still need a bit of an injection of reality. There has been a fundamental shift in the way people are looking at and using AI right now, mostly thanks to tools like chat, TBT and co-pilot, but we are still a way off AI truly replicating human thought and behavior, even though the chat box out there sounding pretty conversational these days. What we've focused on at Luminance is a more narrow application of AI, which is designed to change the way that people work, make decisions and ultimately live their lives.

Jaeger Glucina: 8:31

So what do I think things will look like for lawyers in five years time? If you think about it, so much of a lawyer's day-to-day work could be automated. I'm not talking about the really complex work, but the filing, the large-scale document review, due diligence, data subject access requests off the back of GDPR drafting and reviewing low risk but incredibly repetitive contracts like NDAs. If all of that work was automated which I believe it easily, will be in the next five years, because the technology is already there now. Lawyers would get a huge chunk of their day back for creative thinking. So having time to think is when people start to come up with new ideas, which is ultimately what will take society to the next level, and that's what I believe will call the AI revolution.

Andrew Grill: 9:15

I'm glad you mentioned critical thinking, because I think that's a skill a lot of school leavers aren't taught and we'll have more time to think, and I bring up this topic a lot in the AI-related podcast, in fact, susie Allegra, one of my or my two-time guests. She talks about the freedom to think, not just because of human rights, allowing digital law to permeate the new digital space, but also we need that time to think. So I think if we are going to have some of these menial tasks cleaned up and accelerated by AI, we can then have those new ideas and maybe there are creative ways of solving a case, or it can actually look at precedents we hadn't thought about, because we can't read every piece of law that's ever been written, but AI can and it sounds like it has. Do you think there are areas of legal practice that can't be replaced by AI and will still need the dear old lawyer or barrister or solicitor to be in there?

Jaeger Glucina: 10:06

I do, but it's not so much practice areas or particular specialisms within the legal practice that I think can't be replaced by AI. I think it's actually that creative, strategic side of being a lawyer, but also the empathy and emotional intelligence which actually all of the best lawyers have, because they need to understand their clients, they need to know what looks good for their clients and then make a plan around it, so they also need to be able to tell a very, very convincing story that plays on emotion, not just in front of a jury and court, but in many, many different contexts.

Andrew Grill: 10:38

There's a recent case in the US where a new lawyer cited a non-existent case law for in front of a judge. I think he got a lot of trouble. So should lawyers be using journalists AI platforms at all?

Jaeger Glucina: 10:48

I'm pretty sure every lawyer around the world visibly wins when that story broke and I'm also pretty sure no one will be turning to chat to BT for case law in future. But as with any sensitive or really specialist subjects, you just cannot rely solely on generalist AI. It's just not their purpose to be accurate in a legal matter and, worse, they'll give an answer for the sake of it, whether right or wrong. But going back to that well read friend at the dinner party or that very enthusiastic intern, you know, would you want them as your lawyer? Probably not right. But I wouldn't hazard lawyers totally against using generalist AI platforms ever. I mean sure they could help, you know, do better drafting or quicker drafting of really basic things like client engagement letters. They just shouldn't use it for the important work at the end of the day.

Andrew Grill: 11:35

So you've been with luminance from the early years. So how have you seen AI technology evolve and also how have you seen the legal industry embrace AI over this period?

Jaeger Glucina: 11:44

It's been really exciting to see the leaps forward being made in the tech space. Ai is now, you know, for the first time I think I mentioned this earlier really in the public consciousness. And that includes lawyers. I've been, you know, working with them for many years and I saw them at the very beginning, when I started at Luminance, very resistant and skeptical towards AI, very conservative. But now there is a tangible change in that attitude. They're actually quite excited to use and embrace it, and I think that is because they've actually seen tangible benefits for themselves and no longer fear that it's going to take their jobs away.

Jaeger Glucina: 12:21

I've had customers tell me personally that their eyes were literally burning from reading 100 page contracts every day very boring contracts before they started using luminance to help them with that. But others are saving significant amounts of money totally automating certain tasks like NDA review and also leaving the office earlier at night. So at the end of the day, it's the results that will change mindsets and that's where the difference has been seen over the past couple of years. We have two customers that I could speak about, actually, because they're also very vocal about this, and that's Koch Industries and Hitachi. Both have since using luminance, given total autonomy to non legal teams maybe it's commercial teams or even HR to work on certain contract types. That might sound like a small thing, but NDA is often one of the very biggest causes of friction and delay in businesses getting deals done because there are just so many of them. So removing that from the legal team's remit can literally be the thing that reduces a sales cycle and gets deals done quicker.

Andrew Grill: 13:22

I know that luminance focuses exclusively on the legal industry and your LLM has been trained for that, but what other industry could benefit from your experience with the legal industry in adopting business grade AI?

Jaeger Glucina: 13:32

To be honest, any business that deals with contracts on a day to day basis and that's not even an industry specific problem that's essentially every company in the world there are many variations of the quote.

Andrew Grill: 13:43

Ai won't take a job, but someone who knows how to use it will. So how do we close the AI skills gap in the legal industry and more broadly?

Jaeger Glucina: 13:50

I think what's great to see is that lawyers are keener than ever to learn how to get the most out of AI certainly from what I've seen, but I would like to see law firms and businesses make the most of that while they can and mandate the use of AI for certain tasks where it makes sense.

Jaeger Glucina: 14:07

So that's what will force lawyers to change habits, something they're not known to be particularly fond of. Beyond the legal profession, I think reskilling is a method that's going to be key to future-proofing workforces already in technical fields like software development, data science, engineering. Even Then, offering the staff the opportunity to upskill really fosters that environment of continued learning, as well as equipping team members with the skills they need to ensure that their business stays competitive in the face of what is very rapid digital transformation right now and then I think. Finally, increasing capacity for mass engineering and technical education in schools is another way to help establish strong fundamentals that can foster an interest and familiarity with technologies like AI. Company-sponsored AI innovation hubs and labs strategically located near universities. That could also provide an opportunity to tap into emerging talent and provide more of that specialist training while they're still in university or even just recently graduated.

Andrew Grill: 15:06

Now touch before on the difference in the way that lawyers are charging. It's generally an hourly rate, so how might the use of AI change the way lawyers and other professional services firms charge their clients, moving from this hourly rate to more of a project fee, given some AI tools such as yours can cut the effort from hours to seconds.

Jaeger Glucina: 15:22

I've seen many forward-thinking law firms and other professional services firms change up their billing models where offering a technology-assisted service. That was starting to happen even six years ago. So applying a fixed fee model helps them to pitch more competitively for work and also retains the margins, which are, of course, the important part, because they're able to complete the work in a fraction of the time. The added bonus for the firms who've been doing that and their clients is that they're doing better quality work. The reviews are more thorough, meaning advice is therefore more informed and clients can make better decisions.

Jaeger Glucina: 15:58

Happy clients, of course, equal repeat business. So I'd be surprised if more law firms didn't adopt that charging methodology in future. It will probably just take some of the bigger players adopting that approach to create a bit of a ripple effect across the industry, but up until now it's been mostly the small and mid-sized firms who've been early adopters of that method, because they have to compete. What I think will really force the change is the fact that in-house legal teams have now started to adopt technology faster than their external law firms. So law firms are eventually going to see less work come their way, and that's probably what's going to be the biggest incentive for them to rethink their strategy.

Andrew Grill: 16:37

So I'm sure if there are lawyers or legal service support people listening to this podcast right now, they're going yeah, andrew, but that will never work here. So what are the barriers you're seeing for lawyers adopting AI?

Jaeger Glucina: 16:48

There can be a perceived lack of trust in AI and given recent events like the lawyer citing non-existent case law, it's not really hard to see why there is that lack of trust. But a lack of training, reskilling and strategy for the technology when it's brought into a business also doesn't help. Lawyers need to be told what to use it for, when, why, how. They won't just change their habits overnight. The technology needs to be integrated into everyday work and, as with everything, results are what will keep them using it. So it's not always easy, I guess, to assess which AI will have the biggest impact on your legal team, especially when there are so many technologies out there today for porting to be AI. So for that reason we sort of tend to provide our shopping list of key questions for business leaders scouting the market.

Jaeger Glucina: 17:42

Those are three things for me. So one, is the AI instantly deployable, ie no lengthy setup or training, so that lawyers don't just get bored and give up early? The best way to flush this out early is to ask that vendor or product to set up a trial within a few hours. Then you'll really see. Second, can it learn any new piece of information, any concept, in any language, in any jurisdiction, and can it do that on the go without needing someone to come in and be paid to train it? And then three has it just slapped on an integration with an app like ChatGPT and then called itself AI, or does it have a heritage and foundation in true AI, with customers that can attest to that?

Andrew Grill: 18:20

I want to touch on regulation, because what ChatGPT did back in November 2022 was it woke up the regulators, the fact that this can actually do a lot of things that they hadn't considered before. So where do you think regulation needs to play in the role of adoption of AI in the legal industry and more broadly?

Jaeger Glucina: 18:37

I think people think that it's quite straightforward to regulate AI because they think AI makes decisions in the same way that a human does. But even the way that we make decisions isn't actually very explainable. Have I asked you the simple question of why you had I don't know what you had for breakfast, but why you had waffles for breakfast this morning? Would you be able to give me a fully reasoned answer on it, or was it just a case of felt like it at the time? So the reality, well, I'll let you answer that, but the reality is that people don't really know how AI works, and AI regulation should start with bringing in the people who are actually experts in the field. Ai in the legal industry presents a pretty unique challenge because of the sensitive nature of the work. That's why we say legal grade, specialist AI is really important.

Andrew Grill: 19:27

Digitally curious is a phrase I use a lot. Actually, it's the title of my upcoming book. So how can lawyers become more digitally curious and, in turn, more digitally literate?

Jaeger Glucina: 19:35

First of all, congratulations on your new book. I'll have to make sure I get one from the store. And from my perspective, lawyers are becoming more digitally curious. One because the technology is becoming more user-friendly and effective, but two because they are starting to have no choice. So a great example is the data subject access requests, or DSARs as they're known to the legal industry, which are pouring in off the back of the GDPR. No legal team could hope to complete a DSAR and be confident that they'd caught everything without the use of technology.

Jaeger Glucina: 20:10

And then, when your business is financial and reputational damage is potentially at play, the stakes just get higher. Tesco is one business that uses luminance to complete DSARs, and you can only imagine how many the likes of NatWest and Coots had over the past couple of months. One of the reasons we offer a free trial of our technology is because it takes the pressure off lawyers. They can play around with the tech and they can get comfortable at their own pace. Over two weeks, Usually we find they actually toss. They love it very quickly, but then so they're going to miss it, while their businesses have to go through the usual procurement steps to bring new technology in. So that would be. My feeling right now is just that they are curious and we've just got to capitalize on that while we have the chance.

Andrew Grill: 20:56

You say that without an adequately prepared workforce, the role out of AI has the potential to stifle innovation and limit businesses' ability to leverage technology. What can we do now to prepare?

Jaeger Glucina: 21:06

To start with, there needs to be a broader social education piece of work done to familiarize people with what AI is and how it will help improve their quality of life. We don't need everyone to achieve a high level of technical understanding. That is totally unrealistic in my view but we should teach people to understand key concepts that underpin AI, like probability and risk. If the public were clued up on what AI can and can't do and the different types of AI for example, the difference between narrow and general AI then the workforce will generally become more comfortable and informed when working alongside AI day to day. We should also make a really concerted effort to encourage underrepresented groups into STEM subjects.

Jaeger Glucina: 21:52

There's a lot of research out there supporting the idea that diverse workforces achieve better results, so that has the potential to not only bolster technological innovation but also address the current AI talent shortage. That work could start at the grassroots level, but with increased investment into basic digital skills programs and venues like libraries that can provide access to technology. That's where you're going to make a difference. Then you can also, as a business, you could partner with universities and research institutions to set up AI innovation labs. That would allow the UK to be more of a key driver of AI innovation but also enable businesses to tap into the emerging talent that is so great in the UK and also provide that really important specialist training early in the careers of the people who are looking to go into those AI fields.

Andrew Grill: 22:41

So I want to go back to your own use of AI. What was the first thing you asked ChatGPT, and do you use it in your daily workflow?

Jaeger Glucina: 22:48

I actually used it to recommend to me the top lead generation software, which is something I was looking into for our commercial team and, to be honest, it is perfect for things like that. I got a great answer and I've now selected a fantastic product. My job is definitely not legal these days. It's much more heavily centered around people. So, to be honest, I don't find that I use it that much day to day. I am always playing around with our chatbot asking me I ask it, you know lots of different creative questions every day to put it through its paces, and it hasn't yet disappointed me. So we shall see.

Andrew Grill: 23:23

And how do you keep your own language models, large language models, updated? I mean laws changing all the time, new cases are coming out. How do you constantly make sure it's up to date and ready for fit for purpose?

Jaeger Glucina: 23:34

Well, that happens organically by way of our customer base. So, you know, we have customers working in 80 different languages in 60 different countries, and you've got this really lovely mix now of not just law firms of all sizes but also all of the big four and a huge number of in-house legal teams, everything from your one person legal team all the way through to you know, your huge manufacturer, and they buy by a natural byproduct of just using it, informing that large language model every single day. But we also have a team of in-house analysts, of course, who are making sure that all of that is very much accurate and consistent. So the two together is very important.

Andrew Grill: 24:16

So I had Peter Voss on the podcast recently to speak about general AI. What's your view on this and are we likely to see any time soon in the legal industry?

Jaeger Glucina: 24:24

For the past, 50 years or so, we've seen AI being applied to very specific tasks. That's called narrow AI. So essentially, we've found one problem for the AI to solve where we're using narrow AI, and that results in applications like chatbots, self-driving vehicles, translation engines, you name it. In our case, the task is legal document processing. Now, artificial general intelligence is a much bigger idea. Here we're talking about a system that can perform any task that a human being can perform. So that is definitely the direction that we're heading in.

Jaeger Glucina: 24:59

But for me, we're quite a few years, if not decades, off AGI, and that's mainly because, if you look at the state of AI research and funding right now, we're still not investing that heavily into the pursuit of AGI, certainly not to the extent we need if we're going to attain it anytime soon. So for the time being, I think narrow AI applications are going to be far more achievable and far more profitable for businesses. Businesses aren't straying too far outside the limits of what their AI can actually do, because AI only has a certain latitude to play with in order to learn how to do a task. When you start asking it to do other things, you'll typically see performance on the first and primary task go down. So ultimately, my answer is that AGI is where we're heading, but it's nothing to be afraid of just yet. The AI overlords are still a little way off. Plus, we have that very interesting question, which is do we actually want to achieve AGI? I'm sure we're going to see plenty of debates over that in the coming years.

Andrew Grill: 25:56

We're almost out of time. We're up to my favourite part of the show, the Quick Fire Round. We learn more about our guests iPhone or Android.

Jaeger Glucina: 26:02

Now it went for iPhone purely because of AirDrop. Android wins for camera Window or aisle Isle. I'm quite tall.

Andrew Grill: 26:08

In the room or in the metaverse, nothing beats in the room. Your biggest hope for this year and next More automation. I wish that AI could do all of my.

Jaeger Glucina: 26:15

Sleeping. I would love to have more hours in the day.

Andrew Grill: 26:18

The app you use most on your phone City Mapper. What's the best advice you've ever received?

Jaeger Glucina: 26:22

Can't expect to learn everything on your own. Seek out the wisdom of others. What are you reading at the moment? I'm reading the third book in the Three Body Problem Trilogy, which is titled Death's End, by Lu Xixin. Who should I invite next on the podcast? I think Sam Altman. I'd be very interested to hear what he has to say about his vision for open AI in the future and the role that he's playing in settling the AI regulatory agenda.

Andrew Grill: 26:46

How do you want to be remembered?

Jaeger Glucina: 26:48

As someone who always gave 200%, whether that's parenting, being a leader or being a core part of the AI revolution for the legal profession.

Andrew Grill: 26:57

As this is the actionable future, as podcast. What three actionable things should our audience do today to prepare for a world of business-grade AI, or what I call enterprise GPT?

Jaeger Glucina: 27:07

First, be curious whether that's getting your hands on chat, gpt or Adobe's generative AI or co-pilot, which is coming soon, or a more specific enterprise solution. Don't be afraid to simply give AI a go, a test and learn. Strategy puts curiosity into action. Once you get experimenting, you'll become much more comfortable with the idea of AI, what it can do and how it can help your business. Second, learn to recognize real intelligence. So, carrying on from the last point, once you start using AI and learn what it can do, you also learn what it can't do.

Jaeger Glucina: 27:42

So it's probably going to become a bit of a tired saying, but not all AI is actually intelligent. So, as it becomes increasingly embedded in our lives, products purporting to be AI are going to capitalize on a lack of AI savviness amongst consumers. If you learn what really constitutes intelligence early on, that's what will mitigate the risk when you're purchasing software or something similar for your business. And then my final point, number three, embrace automation. Nearly every job and workflow can be improved by some degree by automation, and that is not something to be feared. So find the parts of your workflow or your day-to-day life that can be streamlined and recognize the area of your work that could benefit from more time back and more resource. Automation really can be a positive catalyst for business change.

Andrew Grill: 28:36

Three very actionable tips there. Jager a fascinating discussion. How can we find out more about you and your work?

Jaeger Glucina: 28:42

Actually, I still give demos, so get in touch. I'm on LinkedIn, send me an email and I'd be happy to meet up, or come to our website and you'll get to see a little snapshot of what we do.

Andrew Grill: 28:52

Thank you so much for your time.

Intro: 28:53

Thank you Until next time. This has been the actionable futurist podcast.