The robots are coming, but don’t panic: The AI legal revolution explained
Irish Legal News reporter Kevin Burns meets former lawyer Noah Waisberg, co-founder and CEO of Canadian legal tech company Kira Systems, who is in Dublin to meet Irish clients and spread the word that Artificial Intelligence is an opportunity and not a threat – that it should be embraced and not dreaded.
Welcome to Ireland! We understand you’re here to meet with your existing Irish clients McCann FitzGerald and to meet potential clients. What is the core message you’ve been trying to get across to your audience?
Firstly, that incorporating efficiencies into law practices is the direction of travel in a lot of other markets, so it’s likely that this will be the case here as well. There are real advantages to embracing these changes at an early stage of the trend, because it’s not simply the case that you can just buy the technology and immediately catch up, rather it’s the first step towards modernisation in this area. That said, I’m sure there will be a few companies that can start late in the game if they’re particularly well positioned.
My second message is — the future is friendly. This doesn’t have to be some awful thing. Many of the lawyers I’ve met are in fact earning more money as a result of their firms embracing AI technology. The robots aren’t coming in to take your jobs!
“The future is friendly. This doesn’t have to be some awful thing. Many of the lawyers I’ve met are in fact earning more money as a result of their firms embracing AI technology.”
In what way is AI technology currently used in Irish law firms?
AI is predominantly used in three different ways in Ireland. The first is in e-discovery (i.e. discovery in legal proceedings such as litigation, government investigations, or Freedom of Information Act requests, where the information sought is in electronic format). Technology-Assisted Review (TAR) is basically a form of AI, where you have software which help lawyers decide which documents are relevant to a particular given investigation, and eventually determining which documents are submitted as evidence.
The second is Contract Analysis software in use where lawyers are doing M&A (mergers and acquisitions) deals. In the past, you’ve seen huge teams of twenty to sixty lawyers deployed just for reviewing M&A documents and spending maybe 60%, maybe 90%, on their work day going over these documents which could be spent on doing much more productive work. And this software does as accurate, potentially more accurate, work in a fraction of the time.
The final one is Expert Systems. I know that Novologic Inc. and Austere Technology Solutions, two of the big AI companies, are moving into building automated questioning and answering systems for law firms.
Do you have any predictions as to how this is going to change in future?
Well there’s two parts to it. Firstly, you’re going to see more and more options for the customers of law firms. So, in Ireland, you’re going to see more and more market penetration, and that’s been seen in other countries too. In Ireland you have the “Big Five Firms” here, the equivalent in Canada is the “Seven Sisters”. Four of the Seven Sisters are our clients. In Australia it’s the “Big Six”; four of the Big Six are also our clients. In the UK, we’re partnered with Freshfields Bruckhaus Deringer LLP, Clifford Chance LLP, DLA Piper LLP, Addleshaw Goddard LLP, Ashurst LLP, HSS, and more.
I was in London earlier this week and one firm I was meeting with said they’d done some market analysis and 47 of the top 50 firms had announced that they were using some form of AI.
Some of those firms might be over-exaggerating the extent of their AI use, but the majority of them aren’t. So, I think AI is also very likely to proliferate over here. So in the short term, I think that AI is going to spread to law firms in Ireland.
What’s more interesting is the long-term impact of AI. Short-term, firms are going to do deals and discover legal documents more efficiently because of the use of technology. But in the long-term, AI is going to change their business model. It’s a fact that there’s a lot of latent need for law and lawyers that is not being served right now because the price is too high. But there are all sorts of things that firms might be able to do if it was cheaper.
It’s a business model transformation through increased efficiency which doesn’t mean at all that the product is going to be inferior. I can see that happening over the next ten years. But in terms of the next two years, we’re going to see all the big firms using AI technology a lot more.
The more interesting part in the extra benefits that are going to come with AI. People are going to be inventing stuff that we just can’t even think of right now because we haven’t seen the roll-out of AI yet.
The analogy I really like is refrigeration. The average fridge took on average 800 kilowatts per year to run in the ’70s. Today it takes around 200 kilowatts a year to run a fridge that’s also cheaper to buy and costs less to run. So, you’d think that we’d use less electricity for refrigeration now, right? But no! In fact, we use more electricity because we use way more refrigeration. People tend to have more than one fridge-freezer, and grocery stores use so much more today than they used to. Today’s Tesco Express uses as much refrigeration as a big Tesco supermarket used maybe only 20 years ago.
Because of this proliferation of refrigeration, new things become possible. For example, at home in Toronto I can buy these frozen, ready-to-bake croissants, so I can give my kids a fresh croissant for breakfast every day, all I have to do is buy five packs a week and freeze them, and that’s all because of refrigeration. And that’s clearly good; I might buy some more of those croissants, my kids will continue to enjoy them. I might even buy some fresh kale juice - twenty years ago no one would have imagined that you could buy kale juice at the grocery store. But now, because of refrigeration, you can easily imagine it and, in a way, we’re all better off for it. And so, there are cool new kale juice businesses popping up that twenty years ago wouldn’t have made any sense.
“Law firms are going to be able to provide services and deliver value to clients that we just can’t imagine right now.”
So the spread of AI will spread the same sort of opportunities that refrigeration brought to the grocery business to the legal sector?
Exactly! Law firms are going to be able to provide services and deliver value to clients that we just can’t imagine right now. And that’s not all AI; that’s just efficiency broadly. AI is definitely part of that, but there’s also perfecting process and choosing the right people. For example, there are currently London-based law firms that outsource some of their business to companies based in Northern Ireland because they can get the same work done for less money.
McCann FitzGerald is Kira Systems first partner in Ireland, is that correct?
Yes, McCann FitzGerald are our first and only partner in Ireland right now, but we’re optimistic that that is soon to change. We know that there’s a lot of very serious law firms over here, and as a lot of other big law firms elsewhere are using our tech, we hope that more of the firms here will partner with us soon. I suppose there’s a degree to which the Irish firms are competing with firms that aren’t based in Ireland but practice here. For example, some of our clients like DLA Piper are planning to set up shop in the Republic of Ireland. So, while McCann FitzGerald might be the only Irish firm we’re partnered with, there may well be a few international law firms that practice that practice here that we’re also partnered with.
Matheson has put out the idea that the lawyers of the future may need to learn how to code. Do you agree with this?
I don’t think so, actually. I used to be a lawyer, I did mergers and acquisition for Weil, Gotshal & Manges LLP in New York. When I was setting up this business, I ran into some tech problems. Now, it would be nice if lawyers knew how to code - I’m definitely not against that - but I don’t know how to code. The good thing about that was that in order to get the business to work I needed to get a co-founder who is really a great technologist.
Our co-founder is Dr. Alexander Hudek, who got his PhD in Computer Science and Mathematics from the University of Waterloo, which is kind of like Canada’s MIT, it’s a great technical school. I think he got his first computer that he started programming when he was eight. He’s a smart guy. So, what am I going to do? I’ve finished law school; how ambitious do I have to be to learn to code to a point where I can compete with guys like Alex? That guy’s undergraduate was in quantum physics, computer science and heavy maths. There are lawyers who are good at maths, but it’s unlikely that they’re going to be as good as Alex. So, is it really the best investment of your time to try and learn to code like guys like him?
“I think it’s really useful to learn how to code; I look forward to a less busy time in my life when I might try to learn. But I think the business is better off because we worked together within our different specialities.”
What we found in the case of our company is that my co-founder and other technologists we hired have really important skill sets, but we also hire a lot of lawyers too. The lawyers teach the technologists how to build a system that can analyse contracts; they don’t need to learn to code to do that, they just need to be really good lawyers. So that was the case for Alex and I. I have a lot of highly specialised skills that I used to get paid quite a lot to use for a law firm. A lawyer at a high-end firm like Matheson will have a skill-set that clients are willing to pay several hundred euros an hour for their time; so, what are you going to do? Are they going to have to go back to learn how to do the coding work at a level that they might only get paid €15 an hour for? Why not let lawyers do the things that they’re really great at, and figure out how to work with technologists at what they’re really great at.
I want to stress that I think it’s really useful to learn how to code; I look forward to a less busy time in my life when I might try to learn. But I think the business is better off because we worked together within our different specialities.
What has been the attitude of the Irish lawyers you’ve met to the AI tech you are bringing in? Are they scared or excited?
Probably a bit of both. Some seem excited, some might be worried about what they perceive as another threat to their jobs. It’s interesting to see how the younger senior associates view the partners in their firms. They are aware that life was pretty good for their predecessors. If you’re a 50 year-old partner at a law firm you’ve had quite a good career, the legal climate has been good to you, and therefore people with less experience have the desire to attain that same “good life” and are afraid that they’re not going to get it and so they’re worried about that more broadly. Some will realise that the world is just changing and that you can either embrace it or fight it, but fighting it is going to get harder as time goes on.
What about law schools and universities? Are they keeping up pace with this change?
I think law schools are increasingly understanding the pace of change. That said, we do tech for corporate lawyers; they don’t really prepare students to for corporate practice at law school anyway. I think that law schools have been getting better at providing practical training, for example, on Wednesday I met a lot of legal academics in London, and there are definitely schools there where the academics are serious about legal tech. But for law schools as a whole it’s still a work in progress.
To what extent will the Government and regulators like the Law Society and the Bar Council be required to facilitate the “AI Revolution”?
There are interesting questions that they’ll have to work through. In a way it’s already begun, governments are having to regulate security tech that is used in conjunction with lawyers. But here will be interesting questions in future, like as AI can solve some legal problems, it can probably already solve some legal problems better than people can. As you see more and more of that, it might lead to the situation where a person is better of going to a computer system than to a lawyer for some problems. That situation will create some interesting problems for regulators to deal with. That said, some regulators have been very encouraging and open to us, for example, the UK has been very friendly to these changes. So I hope that in Ireland it will be a similar case.
Will we need a new legal framework to regulate this legal technology?
I think in the case of e-discovery tech or contract analysis tech it’s already been used under the supervision of lawyers, so if they don’t think it works, they probably just won’t buy it in the first place. If there’s a situation in which there are actual robot lawyers, which I think we are a long, long way away from, then yes we’d probably need some sort of legal framework.
Final question: are lawyers and paralegals going to be replaced by robots?
Some - if its work that is so boring and repetitive that it seems like a robot could do it. That’s how I got started on this line of business; I was doing work and supervising work that it didn’t seem like someone should be getting paid $400 an hour to do. So I thought, if I’m doing work that people shouldn’t be paid $400 an hour to do, why don’t I build something that will help me make better use of my time. If you’re doing work that seems like it could be automated, it’s going to be automated. That doesn’t mean lawyers as a whole are going to be automated, there is just tonnes of other stuff that they could be doing that’s a better use of their time. But if there’s tasks that are just so simple and repetitive and monotonous that you feel it could be outsourced to India or automated, the likelihood is that it will be automated.