How Artificial Intelligence Will Influence the Future of Legal Practice

Episode Notes

Transcript

Law Technology Now

How Artificial Intelligence Will Influence the Future of Legal Practice (Rebroadcast)

05/03/2016

Male Speaker: Hello listeners, this episode originally aired in May of 2016 and we are rebroadcasting it because it’s about Artificial Intelligence and its future impact on the legal profession, so stay tuned. And if you are crewman David Bowman from Discovery 1, you might want to keep an eye on that airlock. Enjoy the episode.

Intro: You are listening to the Legal Talk Network.

[Music]

Monica Bay: Hello. I am Monica Bay.

Bob Ambrogi: And I am Bob Ambrogi.

Monica Bay: We have been writing about law and technology for more than 30 years.

Bob Ambrogi: That’s right. During that time we have witnessed many changes and innovations.

Monica Bay: Technology is improving the practice of law, helping lawyers deliver their services faster and cheaper.

Bob Ambrogi: Which benefits not only lawyers and their clients, but everyone.

Monica Bay: And moves us closer to the goal of access to justice for all.

Bob Ambrogi: Tune in every month as we explore new legal technology and the people behind the tech.

Monica Bay: Here on Law Technology Now.

Bob Ambrogi: Welcome to Law Technology Now. This is your host, Bob Ambrogi. And we are going to be talking today about Artificial Intelligence. And I know sure some of you listeners are saying, oh no, not another program about Artificial Intelligence. It has been all over the news lately, I know. In fact, the cover story of the April issue of the ABA Journal is all about Artificial Intelligence, and that article proclaims that Artificial Intelligence is changing the way lawyers think, the way they do business, and the way they interact with clients. Artificial Intelligence is more than legal technology; it is the next great hope that will revolutionize the legal profession.

But today we are going to talk about maybe a little different angle on this than you have heard about. Back in January I attended an Innovation Summit put on by Thomson Reuters, the company probably most of you best know for Westlaw, and they offered kind of a tantalizing taste of something they are working on — well, various things they are working on involving the cognitive computing power of IBM’s Watson.

You all know Watson as the computer that famously won Jeopardy, and ever since it did win Jeopardy there has been a lot of talk about how Watson might have applicability in the legal field and how it may help to develop Artificial Intelligence products in the legal field.

So today I am going to welcome to our show Eric Laughlin. Eric is the Managing Director of the Corporate Counsel Segment and Legal Managed Services for Thomson Reuters. He in that position is responsible for information software and service offerings, including legal research, regulatory and risk and legal managed services, which includes eDiscovery, contract management and compliance.

He previously worked in the Global Strategy Function for Thomson Legal and Regulatory and has held various marketing roles within the legal business since rejoining Thomson Reuters in 2007.

But of greatest pertinence to today’s topic, Eric is heading up the Watson Initiative at Thomson Reuters, and so we have invited him here today to tell us a little bit more about it.

So welcome to Law Technology Now, Eric Laughlin.

Eric Laughlin: Great. Thank you for welcoming me on the show. I am really excited to be here, and I certainly hope that, yeah, we aren’t just going to repeat everything else that’s been said on AI, but we can bring something new to the conversation.

Bob Ambrogi: Well, I have to say when I attended that summit back in January I was fascinated. I know that Thomson Reuters as a company had announced earlier, last year sometime, that it was working with IBM to develop products on the Watson platform, and that’s throughout Thomson Reuters. I hadn’t really much thought about bringing this to the legal sector. And for a company that has the kind of established credentials and resources of Thomson Reuters to be working on this I thought was really fascinating.

I wonder if we could kind of just start by maybe recapping what you told us at the Innovation Summit. I mean in terms of what it is you are doing. One of the things you said there was that probably the first Thomson Reuters product using Watson that was going to come to market would be in legal. You gave us a little bit of a heads up on that. So maybe you can tell us what you can tell us about that at this point.

Eric Laughlin: Sure. Thomson Reuters is investing in a Center of Excellence for Cognitive Computing, and that’s really an extension of our current R&D efforts into that space. We have a group that’s going to utilize all of the cognitive tools that we have; some that we developed on our own, some that we will partner and sort of license from other folks.

(00:04:50)

And in that vein, the IBM Watson partnership is the first sort of external partner we are bringing into that Center of Excellence. And that’s actually a Thomson Reuters wide initiative, but as you mentioned, the first implementation of a tool that’s powered by Watson will be in legal.

I am working on that project right now and although we are not saying exactly which domains we are pointing it at, it certainly is in the regulatory space. We plan to have something in beta in October and out in the market in the first half of next year.

If you think about — we spend a lot of time already at Thomson sort of embedding parts of cognitive tools into our current solutions, and we can maybe talk about that at another time, Bob, but in addition to what we already have, we are planning to use Watson within a solution for things like a more advanced question and answers, Q&A features, for things like jurisdictional comparisons and for things like tracking changes and sort of recognizing trends in — regulatory trends or enforcement trends.

Bob Ambrogi: I think you did say in January that this would probably be focusing on the global, legal and regulatory requirements as they affect financial services industry perhaps or corporate compliance more broadly. Do I have that much right?

Eric Laughlin: Yeah, the first domain will be more broadly applicable than just financial services; it’s a domain that folks care quite a bit about in every corporate setting. I think whether you misheard or I misspoke at that session but I think people got the perspective that it was just on financial services; it’s more broader than that.

Bob Ambrogi: So let’s talk about definitions a little bit. There is so much talk about Artificial Intelligence, and I am sure in cognitive computing and everybody has their own sense of that. How do you view that? What are you talking about when you are talking about Artificial Intelligence and particularly in the legal space?

Eric Laughlin: Yeah. It’s a harder question than it ought to be. I think definitions are sort of all over the map. It was clear that we have moved from trying to replicate the human brain into taking advantage of a lot of computing power to achieve things that we tend to associate with human thinking. I think there is a crowd of sort of technologists out there that use definitions that frankly are beyond the comprehension of most of us non-technologists.

I think on the other side, the non-techies out there use words like Artificial Intelligence or cognitive tools really to describe anything that seems sort of magical. If an achievement seems sort of out of reach, people say it’s Artificial Intelligence. Once it’s achieved it sort of starts to feel a little more mechanical. So the definitions are hard.

The best I can do is that it’s not a single discipline. I think there are three core elements to it. One is interaction with the physical world. So a computer and physical world interaction, in that I am thinking about things like computer vision or robotics, ways that it interacts with the physical world.

Secondly, interaction with humans. So things like Speech to Text, having a conversation between a user and a machine. Exemplified by I think question and answer sort of interactions.

And then things like knowledge representation, ways that information can be displayed in novel ways, given a deeper understanding of a huge dataset.

So it’s interaction with the physical world, the interaction with a human, and then the underlying technology that really allows learning over large datasets. So you hear things like natural language processing, essentially extracting and classifying text in machine learning.

And I think critically for those three elements, it’s a continuously learning, continuously improving system. And so I think that’s how I think about the sort of multidisciplinary approach to cognitive.

I think there are some interesting visuals out there for our listeners. Michael Mills from Neota Logic has a nice family tree of Artificial Intelligence, so you can look up. You can actually find that in an article he wrote recently for Thomson Reuters’ Legal Executive Institute and probably some other places that Michael has published that family tree.

I think as a community, as a legal community we are sort of enamored with thinking about the technology right now, and I think what all of us would sort of like to see is a shift from trying to describe the technology, to try to describe why and how it can be valuable in the legal ecosystem.

I think that’s inevitable. Something, before it emerges, it’s sort of fun to call it Artificial Intelligence and after it emerges it’s a valuable feature that I find on my research tool or on my phone or things like that. And so you sort of see things starting to emerge in the market that are using a subset of the cognitive tools that are out there.

I would point to things like Expert Systems, which some argue is AI, some argue is not, but really rules-based complex decision trees that are out there; Neota Logic again being one example.

(00:09:58)

Interestingly, Thomson Reuters Legal Managed Services has a partnership with a law firm, with Akerman, and we are supplying sort of tailored regulatory research that’s helping to inform an Expert System that they are making available to their clients. So that’s sort of one place you sort of see an AI application out there already.

Bob Ambrogi: I mean one of the ones that’s getting a lot of attention is ROSS, which is this supposed legal research Artificial Intelligence system developed out of the University of Toronto, now commercialized and moved to Silicon Valley. And the idea, at least as I understand it, I haven’t used ROSS, but the idea is you ask ROSS a question and it spits out the answer. And essentially it’s — the idea being that it’s looking at the law and analyzing your question and applying it to the law and coming up with an answer. And then you could give feedback to the system on whether that answer — on the quality of that answer, the relevance of that answer, the appropriateness of that answer, and ROSS learns from that feedback and supposedly gets better and better.

Does that at all sound like what we might be seeing you guys developing?

Eric Laughlin: Yes. And what’s interesting, we already have some of those tools in Westlaw, and in fact, we use a lot of the sort of underlying cognitive tools already in Westlaw, and you sort of see those things showing up in a couple of different ways.

Things like learning, like sort of continuous learning that we employ in order to learn how to rank the results that come back within Westlaw after a query, which uses machine learning techniques is a feature that we are all familiar with around search of, did you mean this, which uses a lot of sort of character language models.

And then there are some newer features, which actually some are already sort of mistaking for Watson-Powered, things like our research recommendations. We have the ability to, as a user is going through a research session, we can sort of recognize what they are working on from the whole research session, not just from an individual query, but look at the whole research session and suggest the documents that they need to complete their research task.

And then lastly and sort of exactly like the way that you just described ROSS, enter a question and we will give you the answer; we call it Westlaw Answers. So those are things we have already developed. And I think that’s sort of a good lead into our perspective on how to use Artificial Intelligence, which is that Artificial Intelligence is not a product and it’s not the singular thing we should be focused on. Some of the tools and techniques that are represented, the cognitive tools and Artificial Intelligence should be embedded into solutions and really power solutions and evolve them over time.

And so I think you will see a lot of that in our current roadmap for Westlaw, you will see that in our future roadmap as we embed more and more cognitive tools into those solutions.

Bob Ambrogi: So does that mean that you are seeing Artificial Intelligence not as — let’s say, you are seeing Watson not as a platform necessarily for creating new products, but for enhancing existing products?

Eric Laughlin: For enhancing existing products, for creating new features in areas where we sort of might have not been able to gotten to ourselves given our research budget and also in creating certain sort of new products. I am not going to go too far into our roadmap, but if you think about new products —

Bob Ambrogi: Oh, come on, we want to know all about it.

Eric Laughlin: New products that would be really interesting in the legal space and sort of natural extensions from where we are, you think about something like combining an idea like Amazon’s Alexa, that’s always on listening speaker, with a Westlaw Research Spot.

If you have sort of a pervasive listening solution that you could have, Bob, in one of your meetings that could sort of understand the topic that’s being discussed, could do some research in the background and sort of suggest whether it’s doing some issue spotting or some sort of research on the fly for you, that’s sort of a contextual approach to research that is very different than keying in a query or keying in a search string.

Bob Ambrogi: And if it could order pizza for the meeting, all the better.

Eric Laughlin: Absolutely! Exactly! And put on your favorite Prince’s tunes, since we are all mourning his death today.

Bob Ambrogi: Yes, yes, yeah. So is Thomson Reuters going down this road — to what extent are your customers driving or influencing your decision to go down this road and focus on cognitive computing this way?

Eric Laughlin: Well, our customers actually had a really positive reaction to the announcement, and when I have sort of been out talking with our clients, the reaction has been uniformly positive.

(00:14:53)

I think there’s a lot of debate and sort of fear sort of pervading some of the blogs and some of the sort of market chatter around sort of AI robots and things like that, but in the conversations that I have had with our clients, I think what they have recognized over the last year or two of exploring potential cognitive solutions is that it’s quite expensive, it’s quite time-consuming, and the majority of law firms or corporate legal departments simply don’t have the bandwidth to tackle massive AI investment on their own.

And so the feedback that I have been getting is that it’s great that Thomson Reuters is actually instead of — it’s trying to kind of cut through the talk and actually do something and try to prove out where AI will actually make sense and actually have value to them, and then they can piggyback on that investment.

And so there are some firms out there that are thinking about new ways of tackling knowledge management; if old models aren’t working, can Artificial Intelligence sort of, added to what Thomson Reuters is doing, can they then add their own knowledge and their own training to our tool in order to enhance it and make it sort of specific to their firm.

There are some firms that are thinking about using cognitive tools as a way to make money, and I think they also recognize that Thomson Reuters investment in training and content will sort of speed up what they are able to do with that.

There is a senior partner at a Magic Circle firm that put it this way to me, he said, we want to be involved on the ground floor on these things so that we don’t wake up one day and sort of AI has changed our business. We want to be part of that change. And so I think they are coming to us as a potential partner to help them in that evolution.

Bob Ambrogi: There is that fear, Eric, you just hit on the elephant in the room whenever this discussion comes up. There was just a conference this month at Vanderbilt Law School on Artificial Intelligence that had the title ‘Watson, Esq.: Will Your Next Lawyer Be a Machine?’ What about that fear that Artificial Intelligence means a future of robo lawyers, how legitimate is that?

Eric Laughlin: I think the fear is out there. I don’t place a lot of — I don’t spend a lot of my hours worrying about that myself. I think in the short and sort of medium term these solutions will help augment the intelligence of the lawyers already in their jobs and trying to advise clients. I think it will allow them to get through research tasks much more quickly to make more informed judgments on behalf of their clients.

But I think it’s also real to say that there are things that lawyers do today that they won’t have to do in the future, and I think that that’s real. I don’t know — I guess I can’t really be the judge on which school of thought wins out. Does that mean that there will be more work created and lawyers will be able to do that, or does that mean that a lot of the tasks that lawyers do will now go to machines or non-lawyers? I guess I lean towards the idea that if you break down the tasks that lawyers do everyday, there are more and more of them that will be handled either by non-lawyers plus machines approach or the machine itself. So it does feel like that’s an inevitable sort of breakdown.

Bob Ambrogi: Yeah. But one that makes sense, I think, one that makes a lot of sense.

Eric Laughlin: Yeah.

Bob Ambrogi: And the whole idea of machine learning is that machine learning requires human and machine interaction; it’s not that the machine operates entirely on its own without that interaction and that ability to learn. So I think as you say, ideally it’s a way to amplify the ability of lawyers to do their work and to do their job, not to remove them from doing it.

Eric Laughlin: Right. And I think one sort of interesting element to that that I have just started sort of thinking about is, what does this do to the training and sort of building of the skill set for a junior lawyer? If the old model was, a lawyer comes out of law school and joins a firm, does a lot of grunt work in the first few years to not only sort of learn how to research, but learn how to think like a lawyer and learn how to really work for that firm and for a client; that model may be shifting more and more to lawyers going straight to in-house counsel, where they don’t get the first couple of years of law firm training.

And then over time may shift to a point where they are not actually doing any of that grunt work because it is so well-informed by Artificial Intelligence. And in that world, how do we train our next generation of legal leaders if that’s not the experience they have had. And I am not saying that that’s impossible; I am just saying it’s something we need to tackle as a legal industry.

Bob Ambrogi: Yeah. There was an article in ‘The New York Times’ earlier this month that talked about some of the challenges companies are facing in attempting to use the Watson platform to develop applications. I think it focused more on the healthcare industry and some of the challenges that companies in that industry have found, and found it perhaps much harder than they expected.

What kinds of challenges have you faced or found and encountered in trying to adapt Watson to the legal industry?

(00:20:06)

Eric Laughlin: Yeah, I haven’t read that article, but I can certainly speak from our experience over the past few months, a few sort of categories there. One is the investment required in both content and training is material. You need to have, and we have had to, not only bring our content to the table, which was sort of the obvious sort of synergy between IBM and Thomson Reuters, but also bring a lot of manpower and expertise.

As you can imagine, it’s sort of garbage in garbage out situation. If you don’t train the system with experts, then the system will not be expert. So we have spent a lot of bandwidth and sort of investment in making sure that we have the right people doing the training on the system. So that’s one challenge that’s very real.

The other is that it’s not enough to have a piece of technology and then have subject matter experts. You really need to have sort of a solution design. Talented people that think about solution design and respect and understand the subject matter, on one hand, but also deeply understand the opportunities that these algorithms present. And so that sort of solution design skill set is something that is really important to bring to the table in these cases.

And I think that a lot of organizations, they haven’t developed that muscle over time. I think we are lucky at Thomson Reuters in that we have spent decades developing that muscle of trying to marry up good technology with good content and expertise, but that solution design sort of mindset and talent skill set is something that a lot of organizations might not have right out of the gate. So in order to facilitate that even more that’s one of the reasons that we created this Center of Excellence for Cognitive Computing, not just to help understand the technology, but to really marry up the skill sets in the middle. So I think those are sort of two categories.

Bob Ambrogi: Eric, I think one of the questions that a lot of people wonder about is how do lawyers start to trust AI as a solution for legal research and other applications in the legal field, what do you say to that?

Eric Laughlin: I think it’s a really important question, because I don’t believe that we are at a level of trust right now where a lawyer wants to simply ask a question and get an answer and feel a 100% confident that that’s correct, if it’s AI behind-the-scenes. And so at Thomson we are going to spend a lot of time during our run up to launch and in our beta period working with users over and over again to get the user experience right, because I think we have to show that when we answer a question that there’s backup, that there’s a reason that we are confident in that answer, because I think showing our work is an important element to trust.

I think over time as well, I think over time folks will get sort of used to that and sort of maybe the need to share your work will lessen over time, but it’s going to be really important right out of the gate.

I think the other thing for us is that we acknowledge that at Thomson Reuters our sort of barrier, our bar is very high. So we are not going to launch a solution where we are 50% confident in an answer and then ask our clients to train it up the rest of the way. We have a much higher bar to set. So we will do a lot of that sort of training and things like that that maybe a startup might not do, where they can sort of release a product and sort of train it over time.

Bob Ambrogi: Yeah. Well, that was certainly one of the things that I thought when I first heard about this was that what makes it interesting that Thomson Reuters is developing AI products is that Thomson Reuters is an established player with a strong reputation and a long track record and that brings a degree of legitimacy to AI products that, as you say, that some startups, they may be fully legitimate startups and brilliant people working at them, but they don’t have that reputation, and for lawyers to trust something I think that’s a key element of it.

Eric Laughlin: At Thomas Reuters the thing that we value the most is the trust that our clients have in us. And so you are absolutely right, we need to uphold that trust as we roll out AI solutions.

Bob Ambrogi: Eric, that ABA Journal article that I mentioned at the beginning of the show, cover story in this month’s ABA Journal, described AI as the next great hope that will revolutionize the legal profession. Do you agree with that? Do you foresee that AI will revolutionize the practice of law?

Eric Laughlin: I think over time it will. I think that that will, as some people are fond of saying, sort of go slowly and then be very, very fast in the end.

(00:24:48)

I think it will change business models for law firms. I think it will inevitably change the dynamic of legal departments, depending on law firms for all the things that they do today. And I think that it will allow legal advice to be more accessible by the masses, I think that will inevitably happen. I don’t know the time scales that we are talking on, but I do believe that that’s likely.

I think that it won’t feel as dramatic as the word sort of revolution sounds, because I strongly believe that it will sort of incrementally appear over time and we will just get used to it and adjust along the way.

I think, sort of back to our consumer lives, if you think about, silly example, but when I get in my car now my phone says, Eric, it looks like you are headed home now. It’s 24 minutes based on current traffic for you to get there. I didn’t ask it anything; it just sort of appeared in my pocket. I didn’t know that that was a feature that the phone was going to add, and I didn’t need to interact with it in any way for it to know that.

And so that’s a very advanced feature of my phone that’s based on all the context that it has about me and my everyday sort of life, and I love it. It’s a great value-added feature. It doesn’t scare me, but it just appeared in my life and I liked it, and I think that’s going to happen over and over again in legal.

I think we are more likely to add features to existing solutions to have new sort of incremental adjustments over time, and I think that people will sort of appreciate them over time, and then you sort of wake up after 10 years and the world does look pretty different.

Bob Ambrogi: Well, I for one am looking forward to the day my Alexa has a Westlaw research bot in it, and until then I really want to thank you for taking the time to talk to us today. This has really been interesting. So Eric, thanks for being with us on Law Technology Now today.

Eric Laughlin: It’s been my pleasure. Thanks Bob.

Bob Ambrogi: Well, that’s it for this episode of Law Technology Now. We have been talking with Eric Laughlin, Managing Director of the Corporate Counsel Segment and Legal Managed Services for Thomson Reuters, and also head of its Watson Initiative. Thanks again to him for joining us today. Please join us next time for another episode of Law Technology Now. This is your host, Bob Ambrogi.

[Music]

Outro: If you would like more information about what you have heard today, please visit  HYPERLINK “http://www.legaltalknetwork.com/”legaltalknetwork.com. Subscribe via iTunes and RSS, find us on Twitter and Facebook, or download our free Legal Talk Network App in Google Play and iTunes.

The views expressed by the participants of this program are their own and do not represent the views of, nor are they endorsed by Legal Talk Network, its officers, directors, employees, agents, representatives, shareholders and subsidiaries. None of the content should be considered legal advice. As always, consult a lawyer.

Continue Reading