Summer Re-Post: ‘AI + The Legal Renaissance’

[ This is an automated re-post chosen by the AL Editor Algorithm, the original story is from 10th February 2017. ]

When AI first reached the ears of the legal market some years ago there was a flurry of stories about the end of lawyers. For years afterward and with Pavlov dogs-like automation any mention of legal AI summoned up the panicked refrain: ‘The end of lawyers is coming, the end of lawyers is coming!’

This was until law firms and corporates actually started to make use of legal AI systems, especially in the last two years and even more so last year. The clichéd refrain, now exposed to the cleansing light of real experience, seemed to die away upon contact.

It turns out there were no androids or already out of date screen grabs from the 2004 Will Smith movie ‘iRobot’ based on the late great Asimov novel. There was just code, and it was good code too.

As to the way that AI would ‘disrupt’ the legal industry like a chainsaw cutting through a block of butter? It has changed things, yes, but disrupt? I don’t see any law firms collapsing because of AI. The big law firms that have collapsed recently did so because of weak management and careless accounting, which were very human problems. Legal tech was not the culprit.

In fact, here is a bet: the law firms that are adopting AI systems now will be the firms in five years that have larger revenues, higher profits and more contented staff.

Let Me Describe the Ways

As of today there are many leading law firms around the world making use of legal AI systems. For example, using natural language processing (NLP) and machine learning to crack open multi-page legal documents and decipher their meaning for due diligence, lease extraction and compliance checking work.

Or, AI systems scouring an entire nation’s case law in a few seconds to tell you what litigation strategy may be best in a certain court. Or, plunging into the bottomless pit of an entire practice area’s legal information and returning with useful, actionable insights.

If we add in the expert systems produced by companies such as Neota Logic, then we can count many more instances. And that’s before we start to venture into the territory of the top end of eDiscovery software and the vendors supplying predictive coding systems based on NLP and machine learning.

And then there are the legal AI systems that many non-lawyers in companies use in areas such as procurement and sales contracts, in part so employees don’t need to keep going back to the over-stretched inhouse team. And then there’s the legal AI systems that allow individuals to gain access to justice, via provision of legal knowledge and guidance, or that help the user to complete legal forms without the immediate input of lawyers.

Yet, the legal world has not ended, it’s not even been bruised. There are today more lawyers in the world than there were last year and there will be more again next year. Revenues of law firms are still rising. Clients seem to have ever-greater, more novel and more complex needs that require legal input, whether this is via inhouse lawyers or external firms.

If this was meant to be the beginning of the end of the legal world then it looks like we’ll survive it. It’s probably fair to say that Donald J. Trump’s and the Daily Mail’s continued attacks on the independence of the judiciary are a greater threat to the legal world than AI.

But, even if we have finally sent the flaky futurists off to find new conference crowds to beguile with their apocalyptic visions, then surely there must have been some job losses? At least one or two? It may be early days, but surely some lawyers must have become redundant?

The truth is that as of today there has not been a single job loss because of the use of legal AI systems, not that Artificial Lawyer has heard of, not yet anyway. Zero, zilch, nada…not one. And no firm interviewed has even thought of doing such a thing.

But surely this is illogical? If software can conduct legal work, even if that is ‘process’ work, then surely this has reduced the total hours needed to be worked by lawyers? I.e. isn’t this a zero sum game and if an AI system converts X hours of legal work into a few minutes of processing, then eventually over several months this will mean a lawyer must lose their job?

But this isn’t happening.

The Paralegal’s Productivity Paradox

It is perhaps a sobering thought for anti-AI doom-mongers that increased productivity does not automatically result in the end of white collar workers, at least not higher value professionals such as lawyers.

What has happened so far is that those in law firms wearing the white collars and white blouses are moving up the value chain more quickly, not being defenestrated.

And, Artificial Lawyer has actual proof of this. After a very interesting conversation with a global law firm which is using legal AI systems, it was communicated that the firm’s large teams of paralegals are now increasingly asked to do higher value work. Meanwhile one of the City’s top M&A firms added that AI used in due diligence meant associates could focus on more interesting and also more remunerative work.

For the firm advancing its paralegals some of this higher value work is more like the type of work the junior associates used to do. While some of the associates are now able in turn to work on matters they can earn back their massive salaries for.

After all, did it ever make sense to pay a young person $180,000 a year to mark mistakes with a Stabilo highlighter pen from 8AM to 3AM? (And some law firms still wonder why inhouse lawyers don’t want to pay the hourly rate any longer…?) Would you rent a highlighter pen for $180,000 a year? Stabilo pens are great, but still….

But, hold on a moment, you say, this cannot be possible. We were told, no, we were promised, the end of lawyers. How can AI actually be helping lawyers not only to become more productive, but for it to actually advance their careers?

Sorry to disappoint, but this is what is happening. AI is more a catalyst for a Renaissance of the legal profession than some Dark Ages plague.

AI and the Legal Renaissance

But, someone in the legal world has got to get it? Someone has to pay the price for this new technology? Right? The answer is: yes and no.

This is where the situation gets a bit stranger. The reality is that for people who want to be ‘real lawyers’ i.e. take on the great responsibility of looking after a client’s interests, then the future is bright. And that includes those paralegals who are eager to advance and become part of the profession they’ve dreamed of joining.

But, those paralegals and even those associates, and dare I say it, those salaried partners, who like to just turn up to work, deactivate the brain and go through a long day of relatively well paid, but boring and repetitive activity, for these guys it’s a different story.

Even if a person is qualified as a lawyer, it is their behaviour that marks them out as a real lawyer. As the saying goes: ‘You are what you do.’ And if you underline things all day, everyday, then sorry, but….well, you understand.

And it this group of people who are genuinely at risk from legal AI, because why would you pay someone to do what an AI can do faster, better and cheaper? No one would.

And, why do we have PSLs, or professional support lawyers, when AI research systems can do that work? Moreover, and again sorry, are PSL’s really acting as client advisers? Have they not allowed themselves to slip into the realm of non-lawyer activity that is ripe for AI’s greater efficiency?

LPO staff, who live in a demi-world of half in the legal world and half in a type of data warehouse with PCs rather than forklift trucks, are also at risk. Frankly any large client that is paying dozens of people to break out the Stabilos, whether in London, New York, Paris…or Manchester or Mumbai, are looking at process in an unusual way. Maybe base some of the teams that work with the AI systems in off-shored/near-shored locations, fine, but to rely wholly on repetitive human labour? Where is the value proposition?

Moreover, what is now called ‘Robotic Process Automation’, which means the reduction of many LPO computer-based tasks to more streamlined processes, will also make its impact. (Note: this also has nothing to do with robots.)

But, again, even if you are a qualified lawyer in whatever low cost centre you are based, is repetitive labour that a non-qualified person could do really legal work? Probably not. And, if you have not qualified yet, is this what you want? Probably not. In which case, it’s good to hear that some law firms with process centres are now offering support for non-qualified paralegals to become fully qualified lawyers.

Ultimately, isn’t AI actually going to help lawyers to be real lawyers? And, isn’t that a good thing considering we live in an increasingly complex world that demands people with such skills?


To sum up, the arrival of AI marks a Renaissance for the legal industry because it permits lawyers to be real lawyers again and not tired process units counting down the hours of their day.

After all, isn’t the definition of a lawyer a person who is doing something special in society, i.e. taking on a client’s problems and making it their duty to help them? Isn’t that why membership of the profession is so jealously guarded and so heavily regulated? If this is just any other office job, then why all the fuss to become a lawyer?

But of course, it’s not just any other office job. In which case, maybe AI is the best thing that has happened to lawyers in many decades.

[ Article by Richard Tromans.

Normal AL service will resume 7th August 2018. ]