If the ultimate goal of legal AI is simply to remain ‘in support’ and never really change the status quo, then this entire endeavour will have been in vain. Instead, we should be aiming at replacement. Not of lawyers as a whole – which is impossible – but of as many ‘legal tasks’ as technology permits.
The problem is that so much of the conversation around legal AI is focused on keeping lawyers in exactly the same situation, with them doing the same tasks in mostly the same way, but with a legal AI ‘go faster stripe’ added on, a little AI buddy, a minion, a gopher, to do their bidding. That approach is never going to change the game. (And maybe that’s how a lot of people would like it to remain….?)
This impediment is shaped by something so prevalent that we don’t always see it: the belief that a lawyer should really be ‘the whole means of production’ and that any tech in their purview is simply there to assist as an underling. But, the purpose of legal AI is not to make marginally faster lawyers by carrying them along on a digital palanquin, no more than Henry Ford wanted faster horses.
However – and this is an important caveat – replacement of tasks, or near replacement, does not, despite appearances, equate to replacement of lawyers. There is a key difference and this centres on the nature of discrete tasks as compared to the entire range of activities, both now and potentially in the future, that a lawyer might carry out depending on their circumstances.
Horses, Typists, Lawyers
We didn’t invent the internal combustion engine so that horses could go faster.
We didn’t invent Word Processing software to make typists more versatile.
They were invented to perform a task better than before.
The reason why the horse and cart went out of fashion was because its existence was centred on just one main task: to traverse roads, moving people and goods from A to B. The car did that task better. But, if let’s say horses were capable of a lot, lot more, that was central to our lives, they’d still be with us in large numbers in our cities, just not hauling things from A to B. But, they’re not, so they were ‘task-replaced’.
And as for the typists….well, they didn’t die, they took on new roles, and the task of typing became subsumed into the daily grind of a billion white collar workers all over the world. ‘Typist’ and the ‘task of typing’ were so closely connected that any disruption was fatal. I.e. the job was really just one task, and that’s always a risky position to hold, given that society evolves and technology always advances.
So, what about lawyers? Well, this site – and I’m sure most people – would argue that lawyers are not just ‘one task’ e.g. only redlining, or producing client onboarding letters, or figuring out what a legal bill should be, or arranging documents in a chronological order and so on, even if as junior associates they may spend considerable time on doing the same thing.
There are dozens, if not hundreds, of tasks, under the umbrella term ‘lawyer’. There is also not exactly a singular task known as ‘doing law’, nor is there a new technology that just ‘does all lawyer tasks’ – which it has to be noted, are always evolving as well.
Instead, ‘a lawyer’ is many, many things. Each and every lawyer is a composite of multiple tasks, with those tasks changing as they develop in their career, or change organisation and role.
So, are we trying to get rid of lawyers with legal tech? No. Not at all. In fact, ‘getting rid of lawyers’ is as impossible as ‘getting rid of politicians’ – both are intrinsic aspects of a civilised society. In short, replacing lawyers is a fool’s errand, while replacing as much of a discrete legal task as possible…now that makes sense.
In which case, can the millions of people who own the professional badge of ‘lawyer’ breath a sigh of relief?
Well, not exactly. No.
We need to stop shying away from the fact that legal tech is there to remove tasks (or as much of them as possible) from this composite body, this amalgam of activities. That will result in the job of ‘lawyer’ evolving as well.
And to paraphrase an old saying: ‘It’s time to put away [inefficient] things.’
That will mean overcoming one of the legal sector’s biggest hang-ups: its entirely understandable inward-looking view of anything that changes its place in the world.
Ask Not What Legal Tech Can Do For You…
To paraphrase again, this time with a famous song: ‘You probably think this [legal tech] is about you? Don’t you? Don’t you?’
But it really isn’t.
Here’s another couple of examples from beyond the legal world.
A company develops a much better MRI scanner – it’s really way better than before. It’s faster, goes deeper into the tissue, is more accurate, and in fact, it’s even less expensive than other ones. A doctor whose practice depends on deep tissue analysis really wants to have one.
Why do they want this ‘med-tech’ tool?
Because it will help save lives – it will help their patients, as it can do more for them.
Because it will make their lives a bit easier.
Because it will make the hospital more profits, because it’s cheaper than the old MRI scanner, but they don’t have to tell anyone that and can now charge the same as before, but it actually costs them less.
I’m guessing you picked number 1.
Now a tech company develops a solar panel that is 30% more efficient than any others on the market, providing way more electricity at a much lower cost than before. The national energy company’s engineers want to buy this:
Because they hope to get pay rises by fitting this more efficient ‘energy tool’?
Because they will now have to work less hard fitting these panels, as they are more energy efficient? (In fact, they won’t, they’ll do the same amount of work as before, it will just be more impactful.)
Because they understand that better solar panels are not about them and their immediate needs as engineers, but it’s about creating a more environmental energy system for everyone and this new tech will help.
Hoping you picked 3.
In both examples there is a lot in common. Doctors and engineers are professionals. Their professional goals, i.e. to help people – the former through medicine, the latter via engineering, are both outward-facing.
The tech they use may sometimes make their lives easier, but the end goal for most of it is for their clients. Doctors can save more lives with better MRIs and other tech. Engineers can provide better energy security, build better buildings, and so on, with better tech.
How does this connect to legal tasks? Well, lawyers have become a mass of tasks, so when someone provides tech that takes away the work from that task they can sometimes get defensive.
There is a tendency to view tech as ‘replacing them’, rather than taking on some, or hopefully all one day, of an inefficient, routine task, or perhaps – as with the MRI example – doing that task even better than they could. E.g. there is no way a doctor on their own could do what an MRI scanner can do. It’s truly additive to the possible outcomes. Whereas, few lawyers feel that way about technology; it’s simply a gopher that does what they could do already.
Now, would a doctor object to an AI system replacing their diagnosis skills? Possibly, but then most doctors are not paid by time billed, so maybe an AI diagnosis system – which is only one task out of many that makes up a doctor – would allow them to do more. Maybe reduce risk too. Reduce stress as well…!
The engineer is delighted to have new solar panels, but would they want an AI system to help design the installation diagrams for those panels? Maybe they would feel a threat there. But, maybe they’d also sense that this tech allowed them to take on more jobs? Maybe it – as with the doctors – actually reduced stress and risk?
But, with lawyers – inculcated in a world of zero-sum calculations about time = money and who does what in any work project – they often either see legal AI as supporting them as long as it doesn’t get in their way, or if it looks like it really can do what they can do (especially if billable), then potentially it’s a threat. Imagine if doctors were reticent about using MRI scanners because they believed they were a threat?
To Conclude
We need to do several things with legal AI to get the most out of it, and thinking differently, having a new mindset, is key:
Stop being shy about ‘replacement’. But, not with the goal of replacing lawyers, which is technically impossible, and instead focus on replacing as much of as many tasks as possible so human lawyers don’t have to waste time on those things.
To see legal tech as raising the game, increasing the abilities of lawyers, not just acting as a ‘little helper’ where the lawyer’s centre of attention is all focused on themselves. (i.e. ‘You probably think this song is about you…?’)
To think more broadly about what lawyers are there for and their place in a complex, evolving society and economy – and then how legal tech benefits that society and the economy. I.e. civilisation needs lawyers – and always will – but it doesn’t need them to bill $10,000 to do a task that an AI system can do in 5 seconds with very high accuracy. That is just an unneeded tax on the operation of the economy.
To help lawyers let go of inefficient things. Of course, to let go, even a little, means the tools need to be accurate, useful and dependable. But great progress has been made and changing mindset will be as important as tech improvement.
Richard Tromans, Founder, Artificial Lawyer, May 2025
P.S. I am travelling and back ‘in office’ on Thursday.