Top London solicitor slams ‘arrogant’ Cambridge law students bidding to displace lawyers with AI ‘nonsense’

By on

He didn’t hold back

Cambridge students Ludwig Bull and Jozef Maruscak

An Edwin Coe lawyer has described a Cambridge law student-led ‘lawyer v machine’ challenge as “arrogant nonsense”, before stressing to Legal Cheek that law is “a people to people business”.

The ambitious challenge is the brainchild of Oxbridge lawyers Ludwig Bull, Rebecca Agliolo, Jozef Maruscak and Nadial Abdul, who have been plodding away with designing a lawyer robot for some months now.

First launched as crime-identifying LawBot in October 2016, their service then dabbled in divorce law advice before undergoing a rebrand (LawBot X) and moving onto Facebook messenger. If that wasn’t enough, Bull and friends then announced they were launching a cryptocurrency for the legal market.

Now, the latest update is that the students have renamed LawBot X (it’s now called Elexirr) and are putting on a competition called #ElexirrLawyerChallenge to celebrate.

In a promo video set to some funky music that we’ve embedded below, operations guy Maruscak explained: “we realise that there are many sceptics. That’s why we have decided to put our tech to the ultimate test.”

This test will take the form of a “man v machine” competition in London. A panel will supply factual and legal descriptions of legal problems with a known outcome to a robot team and a team of “England’s brightest legal minds.” Whichever side makes the most correct predictions will win.

Elexirr seems pretty hyped about the whole thing, but one person who doesn’t is David Greene, partner and head of litigation and dispute resolution at London law firm Edwin Coe.

Greene — who readers may remember is representing London School of Economics students in proceedings against the university’s student halls — tweeted:

Speaking to Legal Cheek about the rise in AI/tech in law, he said:

That arrogant toad ‘I told you so!’ Susskind has for long been banging on about technology and its effect on legal services. But his is generally about digital technology and the place it has in making the processing and provision of legal advice and assistance more efficient. That does not counter the fundament of the business being a people to people business.

While Greene does believe the Elexirr challenge is “interesting” and in some ways a decent “pointer” to a future of case-predicting computers, for now machines are still lacking in certain human elements.

Despite Greene’s criticisms, Elexirr is sticking by its tech. Bull told us:

As Greene rightly points out, the impact of AI on law is generally hyped. Talk of doomsday scenarios where machines completely replace lawyers show only a superficial understanding of the problem. Still, the human mind and machines have different qualities and attributes. Elexirr is dedicated to leveraging machine intelligence to do things that humans have difficulties with: e. g. case prediction. We hope to learn more about the law in this way and contribute to the legal profession’s transition into the digital age.

However, Greene still doesn’t seem convinced, particularly with Elexirr’s assertion that: “we don’t believe AI/robots will displace lawyers”. He said: “Of course they will displace lawyers; it may give rise to different work, tech always does but much work will be displaced by AI in time.” The debate continues.

For all the latest commercial awareness info, and advance notification of Legal Cheek’s careers events, sign up to the Legal Cheek Hub.


Also A Journo

For goodness sake, did anyone proofread this before it went out? The mistakes in the direct quotes, the confused dating in the third paragraph, it’s very distracting.



Totally agree, if you’re going to get your baps buttered then do it right!



This article just makes Greene look foolish and out of his depth. All lawyers know you need to know your market especially if you intend on putting your neck out with such statements. Greene where is your PHD in Artifical Intelligence – in what way are you equipped to determine its limits and capabilities? How do you know where things will be in a year never mind a decade on this front. If I was a client and there was a means by which I could process the vast majority of a case without a human I would, it is lawyers and programmers which develop the algorithms remember. You make it sound like it’s a game of pinball where you have to reach a certain score to win the case. Idiotic.


AN Other Anon

Are the people downticking Anonymous’ comment others, like Greene, who feel threatened by this ?



I voted it down because the bot in question is absolutely laughable. It has the legal knowledge of a couple of BPP first year undergrads armed with a Tort Law Nutshells revision book, and the sophistication of the programming is such that I could have put it together in Basic at about the age of 7.

If the whole thing is worth £10 I would be absolutely staggered.


AN Other Anon

Shows a remarkable naivety on your part about how this technology works then.

AI works by having the neural network in case “trained” and learning to learn. It’s not about a direct lookup e.g using a SQL query (“oh damn we forgot to update the Land Law tables.. D’oh..”).

Remember this thing (and I have to say the work is quite impressive) is not Westlaw but it would learn e.g. how to use Westlaw..

Seems that my initial assumption was correct then…



You only need to think for 5 seconds about the idea of a bot ‘learning’ legal principles from a small set of facts and a binary win/lose decision to see how ludicrous it is.

AI works well in games like chess and poker. The rules of the game are relatively simple, but correct strategy is very difficult. A bot can learn correct strategy iteratively by playing millions of games against itself on different settings. It can play against itself *because* it knows the rules of the game before it starts.

What the lawbot is doing is not trying to work out the correct strategy of a game where it knows the rules. It is more akin to trying to work out the rules of the game, where all it has to go on are a few facts and a binary win/lose decision in a set of published cases (perhaps 50,000?). There is nothing to iterate and nothing for a neural network to do. At best the bot can correlate individual facts or combinations of facts with the decision, but once it has done that it is done and there is nothing it can ‘learn’.

Calling whatever it is this bot is doing a neural network is an insult to computer science.

So far as I can tell its creators have never spelled out exactly what the bot is doing (beyond meaningless buzzwords). They also appear never to have explained the basis of their claim that the bot can predict the result of cases correctly 71% of the time.

It wouldn’t greatly surprise me if what they really mean is that the bot was given a set of cases to ‘learn’ from, and it then found a rule that gave the correct decision in that same set of cases 71% of the time, which, of course, involves no element of prediction.

AN Other Anon

Well, it is just a research project but as proofs of concept go, it’s been quite interesting.

Don’t fear the future old bean. It can’t be stopped 🙂

The Original Anon

Interesting comments here – but again I would draw your attention to the concept as is now, and the potential of where it can be in a few years or a decade down the line. Remember a few years ago when Go was considered uncrackable by AI “oh it will never happen, there are just too many computations and options!” Yet a matter of years later that’s another box ticked off. I am saying it is foolish and arrogant to assume much of our function will not be replaced by AI in the future, including large parts of the law. Many of the down voters need to respectfully get their head out of their arse/statute books and look beyond their very immediate horizons.


Shut up Susskind


Machines can’t process client experience, ideosyncracy, and anticipate strategy or vision… which is all part of top flite corporate practice


Act I, Scene i

Client – Hi, I’d like to know how to register a fixed charge over a CREST depository interest and, if that’s possible, whether there might be any hurdles to enforcement further down the line.
AI Lawyer – balls have zero to me to me to me to me to me to me to me to me to
Client – Sorry?
AI Lawyer – balls have a ball to me to me to me to me to me to me to me to meca
Client – You know what, we’re going to try and take a fixed charge anyway and if it doesn’t work, we will sue you. What’s your PI cover again?
*Client sheepishly calls human lawyer* – Geoff, please could you check for me whether you can take a fixed charge of Crest Depository receipts?
Geoff – Sure, I will get one of our NQs to check that for you.
Client – Fine, but remember that on panel rates we don’t pay for your NQ lawyers’ time?



Ah yes, these annoying clients. They seem to forget they need the lawyers because they’ve f@&$ed up or are likely to f@&$ up, and it’s either trainees or NQs that bill them £150~ an hour for basic work, or the partners at £500+ an hour for basic work. On a side note, any lawyer that writes off a single 6 minute unit of chargeable finished product deserves to be struck off for being a coward.


Big boi dolla dolla billz playa US NQ partner man

‘Top lawyer’ and Edwin Coe…?

How many ‘top’ professionals do legal cheek have on the books?!



It’s just CV points for the students

Lawyers have been a part of civilisation for thousands of years – there will still be lawyers working in one form or another, even if the way they work will change



Not in the least concerned about robotic legal services provision. The client knows face to face interaction and relationship building is essential re understanding of their legal problem and explaining and undertaking work to remedy same. Interesting project for the students, but a bit of a distraction from what will be their greatest asset when they become practitioners, namely communication, engendering trust and provision of excellent legal services. Bit of a gimmick really, but not about to tear my hair out!


Tim NicebutDim

True in part, but wouldn’t a client prefer the nice face to face stuff AND a cheaper fee due to the amount of grunt work legal AI can churn through?

Ostrich mentality leaves many by the wayside..



The stuff AI is taking over is very procedural and completely dull. Have you met a conveyancing solicitor? Drying paint has more personality. Also, the majority of the financial incentive will be for the law firms because the clients will get the same fixed fee charged, but there will be no major salary overhead sucking up the profit margins from the volume-based work. If a standard document review that takes a trainee(s) 30-40 hours is charged out at £4,000 normally, that price won’t change. Instead, that firm will make a 99% profit margin on that work if you include the cost of the AI’s 1 second of effort to produce the same work.

Anyways, things like due diligence and document production that was normally done by trainees at all firms will be taken over by AI, and therefor the trainees will likely have more exposure to real client-facing work and providing advice under supervision, such as the trainees/juniors do in Canada and the US (procedural stuff is left to the paralegals and secretaries who will soon be replaced by robots, especially if they can make a good coffee). The rookies may have to do one or two procedural assignments the old fashioned way to get a taste for it, but you won’t have a Clifford Chance trainee making photocopies or highlighting signature lines in a contract for 6 months straight anymore. This may actually make the legal profession fun at the entry level!



It’s a good piece of marketing. Been all over my LinkedIn feed for weeks. As long as they don’t lose hands down in their challenge (which I am assuming is going to have to be on pretty basic parameters) then I am sure they will do well. It is just a tool for referrals for now, which clients may like.
2 issues spring to mind:
1) When this is rolled out, the facts will be presented to the Bot as the client sees them, not necessarily as they are. I don’t think I would have ever lost a case if they were the same.
2) How can it handle nuance and evidence-heavy matters like Family law?



Ironically these morons are only doing this for the betterment of their CVs and the ensuing enhancement of their chances of getting a TC.


AN Other Anon

Why does that make them morons ?

It was a project they were set – they’re having a pretty good stab at it as far as I can see. Wouldn’t you do the same ?



More likely doing it to further their opportunities of receiving funding to become one of thousands of failed silicon roundabout startups.



It depends on the nature of the family data. I once decided to turn down the opportunity to go on a safari holiday. I haven’t really had another opportunity to go since, so I regret my decision massively. I love the idea of seeing majestic creatures out in the wild, rather than behind a case. Nature is beautiful and I think the issue with what these people are trying to do is that everything is put at risk. We should be very careful when lighting matches in petrol stations.



What you fail to see is that it isn’t quite so simple. These so called animals you adore – maybe one day these are no more. They will appear to be there, but it won’t be real. It’ll either be an illusion, or the technology will be there to create the animals that appear to be the same, but inside will be computers that are running some script to ensure your baggage gets to your room at the local Sheraton. Things are not always as they seem.



The only thing that I’d do if I was a child if you were my friends would mean you all of us all of them are just going through you all day.


World (Paediatric Mental) Health Organisation

Didn’t read the article, nor the comments. Just normal 6-12 yo posts.
Cool LC.



I love you



I love the contrast of the old and the new. old lawyers just won’t accept new technology. it’s time for the younger generation to move in and change things



You make it look like Greene contradicts himself.

In his Tweet he wrote, “AI will not displace us.”

You then quote him to say of AI/Robots, “…of course they will displace lawyers”.



If the advent of this nonsense means I can leave work at 6pm rather than slogging away until midnight every sodding day I welcome it.


Bev keenan

I think a robot lawyer could not have made a bigger mess up than the cps lawyers in my brothers murder trial. Good luck to them. There all bloodsuckers anyway. Prices they charge.



Greene is right. This is nonsense. There are cases pending before the Supreme Court. If these arrogant students are so confident of their creation, why wait until September, why not give it a go and predict the result of the Supreme Court cases now, before the court resumes for Michaelmas term? We can then have a laugh when we compare the results in the autumn.


wannabe sound engineer

Slick looking video let down by crap sound! They should have used lavalier mics or a decent directional hypercardiod mic. These wannabe lawyers know nothing about sound engineering.


Nick Crompton

Because it’s so important for lawyers to know stuff about sound engineering.


Comments are closed.