Arbitration Meets AI

What does it mean to prove something when a signature is binary code, a map can be falsified by an algorithm, and your opponent's evidence was drafted by AI?

 In this episode, Richard Nicholas sits down with Bernhard Maier - arbitration lawyer, arbitrator, and King's College cyberlaw lecturer,  to explore how digital evidence is redefining authenticity in disputes. They cover the forensic shift from wet ink originals to e-signatures, the rise of AI-generated submissions that are driving up client costs, and the global digital divide reshaping international arbitration.

 Bernhard also looks ahead: with billions flowing into AI startups, he expects a wave of disputes in courts and arbitration when deals fall apart, systems fail, and someone has to work out who's responsible.

 Whether you're a disputes lawyer, in-house counsel, or legal tech professional, this episode gives you practical insight into where AI is already changing your practice - and where it's heading.

 Subscribe, share with colleagues in disputes and legal tech, and leave us a review. Got a question about AI and evidence? We'd love to hear it.

Transcript

Welcome And Why AI Matters

Richard 0:08

This is Richard Nicholas, and today I'm joined by Bernhard Maier, who's an arbitration expert at the firm, and we're going to be looking at arbitration and AI and what that means in practice. So, Bernhard, you're an arbitration expert, so to what extent is technology relevant to your practice?

Bernhard 0:32

So that's a great question, Richard, and I'm really pleased to be here today on this podcast. I love podcasts and yeah, I was fortunate enough over the last few years and almost, well it's been more than a decade, scarily enough, but I was fortunate enough to work on a number of forgery cases. So cases involving signatures, physical signatures, and allegations that signature or those signatures were forged and not written by the person who purported to have written them. And we worked with a guy called Peter Titel, who unfortunately is no longer with us, but he was the absolute best in his field and I had the pleasure of working with him in more than one case. And the first question he asked me when I came to him with the document and said, look, one of the parties says this is forged, is he said, "Well, what about the photocopier? Do you have the photocopier on which this document was was photocopied? Do you have the original? Do you have the wet ink original?" And those were all questions of a physical nature. And that's changed completely because we have these e-signatures and today everything is digital, it's code, it's binary ones and zeros. So we're lacking that sort of analog element. And another component is maps. I've worked on some international law cases involving maps. And there was something in the paper recently about how these, the maps in relation to Iran were actually falsified and how tools that ,were maps that were available online are potentially not what they seem and to figure that out requires a completely different skill set.

Richard 2:31

Hmm. I know it's interesting. It's an area where AI is no doubt having an impact as well. What other ways do you think legal practice is changing? What do you see happening different these days?

Bernhard 2:51

So, I had a really fascinating discussion with a client the other day, and she said to me that it was great that you no longer had to stare at a blank piece of paper for for days before starting to draft legal advice. And that painful process no longer exists because you can go to Legora, we have Legora, or other firms use Harvey, you can go to those systems and say, do me a first draft. And no matter how bad it is, it's a first draft, so you're taking out that initial difficult, perhaps the most painful step. But on a human level taming that empty page is actually proven to be the most satisfying part of being a writer, an author, or a lawyer. There's an absolutely wonderful book called Attention Span, which I highly recommend to anyone interested in the psychology of these digital devices. And she did a study and assessed that sort of satisfying feeling when you've tamed the blank page and there's nothing quite like it. I'm sure Richard, you've felt that before. I have, and I'm kind of refusing to let that go and I am still trying to tackle that that blank page without Legora, although we mustn't tell Richard Medd, our Managing Partner that, because he will you will have my head.

Richard 4:17

Absolutely. Yes, so it's the challenge that makes it worthwhile. I mean, I've heard the opposite argument actually, which is that there's no such thing as editor's block, whereas there is writer's block when you're confronted with something that's yet to be written. So it's an interesting idea, and I can certainly see that in the field of law. On this whole development of AI, I saw just recently, just this week actually, I think in the news, I think Kirkland said that they'd be, they're intending to spend is it 500 million and developing own AI, their own proprietary large language model. Do you think that's a good thing?

Speaker 1 5:06

It's certainly an interesting development and one that we as competitors but also as lawyers need to watch closely. Now Kirklands of course not the first firm to do that. I think A&O were the first ones who were creating something or investing in Harvey and sort of tailoring Harvey. But it sort of a double-edged sword. Clients are increasingly using AI to prepare first drafts of documents themselves. So that's kind of empowering when clients are are entering the legal lingo without any perhaps formal legal training. but on the other hand, clients are also creating a tremendous amount of AI slop, the sort garbage in, garbage out kind of situation. And we at the firm are seeing this in the employment team in particular. I spoke to one of our colleagues just last week, and he said that in one case he's receiving hundreds of pages from a litigant in a nd so whereas these documents before would be a few pages long, they're now hundreds of pages long with repetition, verbose, and no more precision. And so that's that's driving up costs. And I've heard that from other legal disciplines as well, that it's actually costing clients more for their lawyers to analyse these repetitive documents that are kind of drafted based on probability and an algorithm rather than actual input.

Richard 6:39

Yes. I certainly see that in the on the data side. We certainly see that with DSARs and that sort of thing with people being able to create and therefore creating complaints and data subject access requests much more easily and that's not a problem, that's confined to the UK, of course. Much of your practice is international work. And it's, it'd be interesting to know what do you think, what do you see as the international dimension to sort of the use of large language models? We've seen you mentioned that we've seen at least one firm that's developing its own system. Clients are also developing their own system. We've got in-house lawyers also using their own products and cross-border instructions of the sort that you get. That must bring in a whole extra layer of sort of complexity when it comes to using AI as well, that I thought.

Bernhard 7:44

Absolutely. And the international dimension of large language models in particular is fascinating but also scary. There was a case a few years ago where Kenyans in Nairobi were paid to watch all of the horrible content to calibrate LLMs and see pictures that no human should be forced or paid to see. But the beneficiaries of these algorithms then often sit outside of Africa. So you've got this exacerbation of the digital divide, unlike anything we've ever seen before, where we in broadly speaking the West have huge bandwidths and can download any content within seconds. And then you've got other jurisdictions where people struggle to even turn on the lights, let alone running a complex algorithm, whether for personal use or for professional use. So this concept of the fundamental right to access to the internet, the digital divide, all of these interpersonal issues are really amplified in the international context. And then of course, you've got this idea that the internet was going to be the great leveler and bring democracy to to everyone, but what we're seeing today is instead the the splinternet, where different jurisdictions are seeing completely different versions of the internet curated. Iran was a, even before this current war, was a very sophisticated state in curating the internet for its citizens. So you're seeing some really interesting geopolitical issues.

Richard 9:36

Yes. And I saw that even the Pope has weighed in on this whole idea of the internet being for sort of AI being used for the benefit of humanity rather than benefiting a few in the Western world, for instance. So it's an interesting geopolitical split in terms of how it's going. And as an arbitrator, as well as sort of running arbitration claims, I understand that you are also an arbitrator, so you also arbitrate claims as part of your role. Is that is that right?

Speaker 1 10:20

Yeah, that' s right. And that's perhaps, I mean, I mustn't say this too loud because my other clients might get a bit cross, but it's possibly one of the things I enjoy the most about the job sort of sitting as the decision maker, being presented with the case. I've sat in a few cases now, and they really allow you to see what matters and change the way you operate as an advocate as well and I really like the technology components where I can use some of my knowledge from teaching as well in practice and one of the recent cases I sat in the key issue was whether a certain technology had worked properly or didn't, and who was at fault if it didn't. And that was really interesting because all the the other two arbitrators also had really top-notch tech knowledge, and so we could really talk at a high level about different versions of software and interplay between software and hardware, and it sort of really tested an area of my knowledge that hadn't been tested before, and I think we're gonna see a lot more of that.

Richard 11:31

Yes, I'm sure that's right. Real sort of expertise in sort of a very niche area, but you've got multiple hats effectively, so as well as being an arbitrator and a lawyer, you also teach a course at King's College. So that's I believe as well. It's on Cyberlaw, is that right?

Speaker 1 12:01

Yes, that's right. It's a course called Cyber Law, Big Data, Algorithmic Governance and Democracy. It's a bit of a mouthful and we came up with this title about 10 years ago with my co-lecturers, Jonathan Price, KC from Doughty Street, and Penelope Neville of 20 Essex. And it's been a tremendous success. It's grown steadily from I think initially six students to last year we had about 160 students, 160 master students and it's been wonderful to follow the trajectory of technology and the trajectory of how the law deals with tech and also some of the anthropological developments, because at the end of the day, law, the code that we we abide by as people is anthropology as well and so we we teach everything from how AI changes the way we think to how it changes the way we practice law, to regulating big technology companies.

Richard 13:03

Yes and what's your favorite class as part of that course then?

Speaker 1 13:08

Ooh, that's a difficult one because I really like all of them. But I think if I had to choose, it might be the impact of the internet on humanity, which only entered the syllabus a few years ago. And it was based on a recent study by UCL. It's available online where they sent some subjects, some trial subjects through SOHO. They had to navigate their way around SOHO and they wired them up to test their brain activity. And one test group was allowed to use Google Maps and the other one wasn't a nd they monitored their brain function and the ones the group where they they didn't use Google Maps, the brain was sort of flickering, and you could see lots of activity, whereas the people who were following Google Maps, there was no brain activity. So you could really see how different the brain worked depending on the digital inputs and the interaction with the real world. And of course, with autonomous driving and our reliance, over-reliance on on technology, arguably how that will change. And of course, the the law-makers need to be need to be on top of that.

Richard 14:20

Yes. So it's interesting the idea of switching off your brain when you use technology and actually being entirely reliant on this thing. And that's from a sort of university perspective, I suppose the interesting concept now, of course, is of course all this AI is obviously available and very much used, I imagine, by students. Does do students use AI in your in your classes?

Speaker 1 14:49

Yes, and that's a really interesting question because initially we were as surprised as everyone else by the rapid growth and the prevalence of of the technology. And so I would say the first year after ChatGPT was released, we were sort of trial and error, we weren't really sure. Now we're asking the students to put together a short note as to how they use AI and almost like a paragraph of a manual. And you can really you can really sense it and what's interesting with AI is of course there's no, I mean there are algorithms that can test whether it's used or likely that it's used, but there's a sense, there's a human sort of innate sense when you read something that you just know. You know when the language is in a certain way, when it's sort of it's almost hyper precise without actually saying anything at all. And so that is a real challenge. And it's also been a great leveler because the grades used to fluctuate widely, and you get really high firsts and really low two twos or even a third, and now there's a sort of convergence, how a lot of the papers are ending up being sort of medium two ones, which is quite interesting. But I feel like we're at the beginning of the journey rather than the end, and it kind of goes back to the empty page problem because we want the students to have that sense of, oh, I've tackled the empty page, the blank page, I've written something, struggled, it's sort of the struggle is real. We want the students to learn about the struggle, but at the same time, Kings is offering all master students a course in prompting. So it's a completely new skill set that they have to learn. So you kind of have to strike the balance. You don't want to say, well, you're not allowed to use AI because they will have to use AI in their future jobs. And as lecturers, we're trying to prepare students for the next steps in in their career. So it's really striking that balance between having that satisfying feeling of having tackled the blank page, but also preparing them for their next steps.

Richard 17:04

Yes. And of course it will be part of the role for those students as they go into work and I'm sure you must see some crossover and some overlap really between your work as the lecturer and in practice as well. Is that, do you see any overlaps there as well?

Speaker 1 17:22

Yes, certainly. So one of my other favorite classes is called taming the giants, which we had to we we used to call it regulating big tech, but taming the giants was a lot more catchy and kind of more adequate and or appropriate in these days. And there was a recent case in Brazil or an investigation in Brazil where the regulators looking into Google's practices of basically using copyrighted material to train LLMs and questioning that and seeing, well, should you be allowed to profit from copyrighted materials of others? So that is something that we address in the course, that is also something that is coming to the fore in litigation, for example.

Richard 18:09

Yes, so work mirroring what you're seeing in academia. No, that's interesting and in terms of sort of wider trends, what are you seeing in terms of AI and disputes? Because you're sort of at the cutting edge. You're seeing, you're on top of the technology, but you're also dealing with the disputes themselves. What are you seeing in terms of those sort of wider trends?

Speaker 1 18:34

Well, I mean, there are so many, so I could go on about it for hours, I won't, I'll try not to bore you. But I think one of the biggest macro trends is how investors at the moment are plowing billions into AI systems. I mean, the numbers, the figures are just astonishing. And the key question will be will these systems work? Will these startups ever make a profit? What if they don't? You're seeing the big fight between OpenAI and Elon Musk spilling out into court. Who will be responsible for failures? I can see a huge fallout coming our way with disputes occurring left, right, and center in courts in arbitration. Take Kirkland, for example, you mentioned before. Law firms are currently charging billions to manage the M&A activity in this space, drafting millions of pages of contracts, data rooms, reviewing and if any one of these deals falls apart, there will be huge disputes and and people fighting over what happened, tracing it back, taking the IPOs as well. I thought the the SpaceX IPO is a really interesting example, and I haven't read the prospectus yet, I'm not sure if it's out, but when I think back to to WeWork, which was a huge corporate failure, again dealing, leading to claims left, right, and center. Adam Newman at the time had come up with this concept of community adjusted EBITDA. So he put in the risk factors, he just came up with the new financial metric and said, Oh, this this is a risk factor. And I think with SpaceX, there will be a huge amount of risk factors. And then Elon Musk being one of the most creative people in this business, he will come up with some seriously interesting language. And then language being the the lifeblood of lawyers, we will have to pour over these risk factors later if there's a fallout and come down one way or another. So I think this will keep us busy for for a very long time.

Richard 20:49

I'm sure you're right. Yeah, so plenty of work still out there for lawyers, for litigators, for those involved in disputes. That's been really interesting. Thank you, Bernhard. Conscious we need to bring this to a close, but in terms of , for anyone who's dealing with disputes, Bernhard is very much the person to come and speak to, whether that's in the university or in an arbitration and people will find you, well people can find your details on the website. So do, are people, are you happy to take calls from people who'd like to know more about what you can do for them?

Bernhard 21:37

Of course, absolutely, Richard. Thank you so much.

Richard 21:40

Excellent. Thanks very much, and look forward to catching up soon.