Summer Re-Post: ‘Legal AI + The Industrialisation of Cognition’

[ This is an automated re-post chosen by the AL Editor Algorithm, the original story is from August 2017. ]

The Industrialisation of Cognition

After the financial crisis of 2008/9, the term ‘the New Normal’ was used to describe how corporate legal teams were suddenly pushing back on bills and demanding transparency, efficiency and project planning.

After years of clients rarely saying no, with few really doubting the external firms’ ‘means of production’, something finally happened. ‘This is not good value,’ the inhouse lawyers dared to utter.

What had changed? They had been pushed internally by the C-suite, whose job it is to do the best for customers and shareholders. They had been asked to justify their legal spend. And, in truth, they couldn’t.

This simple, yet seismic change, paved the way for the eventual entrance of legal AI. As revolutionary leaders will tell you, having a plan to topple the old regime is all fine and good, but you still need ‘the right conditions’ for real change to take place.

And let’s face it, not all inhouse lawyers wanted to push back. The buyer and the seller had become a little too close for comfort, one might say, and it really wasn’t doing the end client – the shareholders of the business – much good.

The surprising thing is that it took something as bad as the worst economic crisis since 1929 for the people who actually owned the businesses that were being served so inefficiently, to realise just how bad a deal they were getting.

And, worse still, it was their own inhouse legal teams that had allowed this to get so out of hand. What kind of repeat buyer lets the seller push the price up every year for doing the same thing in the same way, where the product stays the same and there is no improvement in the outcome?

One answer to that is: a buyer who is faced by a monopoly. But, law is a free market, right? Buyers can pick and choose between several equally good firms? Everyone is always saying how hard it is to differentiate between law firms in each market segment. What was stopping a buyer exercising their power and just picking a firm that was willing to play ball?

And they did. But, few in the legal industry wanted this change. Let’s not kid ourselves here. There were no lawyers demonstrating outside court houses demanding they be allowed to be more efficient. GCs were not resigning from their posts because external law firms had yanked up prices for the 10th year running for the same outcome.

It was driven from outside by market forces and increasing transparency inside companies about how legal work was produced.

Legal sector commentators were fast to see the new trend and declared it to be a new norm, from which there was to be no going back. And they were right there. Though, the higher value, more complex and bespoke advisory matters continue to see increasing charge out rates.

But, for the rest of what a law firm did? There were no more guarantees. A new approach was truly emerging after years of conference discussion and occasional warnings from firebrand GCs that had previously come to little or nothing.

And the tool that would make this leap from artisanal ideas to modern industrial design happen has turned out to be AI.

That said, at first it looked like lawyers might just get away with a bit of outsourcing and wage arbitrage, post-2009. I.e. have nearly the same levels of manual inefficiency, add in a bit of process to neaten things up, add in some people who are paid far less and have less opportunities to develop their careers and bingo – we’ve done it.

But, nope. That wasn’t going to cut it.

That said, it’s not surprising that the legal sector’s first response for what was essentially a demand for industrialisation, was to reach for an equally artisanal and manual approach, in this case ‘legal process centres’.

But this was just the old ways repackaged. ‘Work’ still meant throwing as many bodies as you could at something until it was done. Done more cheaply, yes, primarily to protect the law firm’s own margins that were now under pressure, but that was all. This was not a modern approach, this was like an agrarian society recruiting additional seasonal workers to toil in the fields at harvest time and then laying them off once the crops were collected. All with very little tech involved.

But, this isn’t the 18th century, it’s the 21st.

There Has to Be A Better Way

In a world where clients implicitly demand the industrialisation of cognition, a term Artificial Lawyer developed back in 2016, then legal AI is both essential and timely. That’s not to say GCs knew they wanted AI back in 2009, though many do now. It has just turned out to be the logical answer to the question: ‘How can this be done in a better way?’

But what is meant by the industrialisation of cognition?

The law is a cognitive discipline, it’s all about consuming written information, sorting it, analysing it, extracting key meanings from it and then responding, usually again in written form. Often this cognitive exercise revolves around contracts, the crystalisation of the to-and-fro between two parties.

While some of this is at present too complex for legal AI to industrialise, a large chunk of the simpler and more repetitive tasks are indeed wide open to cognitive industrialisation. The classic example, of course, is document review using machine learning and natural language processing, though there are several others.

It makes sense in the same way that electrically-powered lathes started to be used in factories at the start of the 20th century to produce parts that didn’t really merit hand-crafting anymore. Or, when movable type printing changed to modern electronic printing.

In this way the movement of AI into the legal world, taking on cognitive tasks such as reading repetitive documents and doing ‘the donkey work’, is a ‘new normal’. To be averse to this technological and industrial evolution in the legal world would be to object to the Industrial Revolution and all the gains it has delivered.

Some people have tried to go back to the Golden Age before industrialisation. These include, in no particular order: the Luddites who objected to weaving machinery that was spoiling their artisanal stranglehold on the wool industry; the Khmer Rouge of Cambodia who wanted to take the country back to an agrarian idyll – it didn’t work out well; and (sorry to some old friends) hippies/crusties/New Agers/anti-growth campaigners who think that ‘technology’ is dehumanising and if we could only all live in a pre-industrial world then everything would be better (though they all seem to own the latest products from Apple for some reason….).

The irony of the agrarian and artisanal approach to economics is that although a small number get to live nicely, the system requires a huge number of people to do incredibly dull and unrewarding labour, with little chance of advancement.

Therefore, the arguments for consigning to the history books the process work of paralegals and junior lawyers and replacing it with AI, to Artificial Lawyer at least, seem irrefutable.

In short, legal AI is now here and it’s never leaving the sector. It will steadily spread through law, not like wildfire, but firm by firm, market by market, corporate by corporate and through the LPOs and alternative legal services providers of this world.

Some may predict global legal AI adoption will happen very soon. The reality is that it will not be evenly distributed, just as modernisation of the legal market has been uneven over the last few decades. There will be pools of rapid adoption (see below) and also swathes of law firms that it seems impossible that they have stood still, but they will, often because certain jurisdictions don’t see client demand for change. But, the larger and more international firms will certainly adopt AI at an increasingly rapid rate. Why? Because the clients are up to speed. No choice.

For example, in London, many of the larger law firms are already conducting client work with AI systems involved, or are currently piloting the technology with a view to doing so. This is an extraordinary expansion in adoption, given that commercial uses of legal AI (at least beyond e-discovery) were very limited in the UK prior to 2016.

Leading law firms in the US and Canada are also embracing AI doc review, expert systems, AI driven legal research, litigation prediction and analysis, and more. So too a growing number of firms across mainland Europe. Singapore and India too have joined the movement, so too has Australia. And China has clearly declared it wants to be a pioneer in this space. These markets are adopting legal AI and also in many cases producing companies that supply legal AI technology.

Legal AI’s First Wave, then its Second

To conclude: it’s hard to be efficient if you don’t have the tools to allow you to do so, or the economic imperative to change. Try sawing a tree down with a sharp stick and getting paid by the hour to do it like that; you get the picture.

The other point is that much of the above has focused on lawyers, both inhouse and external, responding to external pressure, rather than leading change. But, as we all know change happens for a reason. And in the past there was no reason to change.

Now, the introduction of an AI toolkit into the legal world, filled with machine learning, NLP and natural language generation, rules based systems, the ability to construct complex predictive algorithms and much more, opens up many opportunities for law firms, clients, LPOs and ALSPs to create solutions that can do far more than tackle inefficiency.

One might say that in commercial terms the second wave of legal AI will be making more value-creating uses from AI, not just doing what obviously no human should have to endure (even if they’re well paid).

The move to combine legal AI applications with expert systems and collaborative technology is a sign of where the market is going. Lawyers are working with tech companies to create holistic solutions to legal work problems and helping clients to do things they may not even have attempted in the past, not just sponging up the lowest levels of process work.

In short, AI handling the lowest rung of process work is just the beginning. We are at the beginning of a remarkable change in the legal market that will be shaped by the industrialisation of cognition. And the best part is, it’s only just begun.


[ Article by Richard Tromans.

Normal AL service will resume 7th August 2018. ]