Hacker News Reader

Comment by nickelcitymario | Hacker News Reader

Having worked in a law office, I can say confidently that the future of legal AI is human+AI. AI can rapidly spot all the issues, saving expensive lawyers from spending their time wading through the documents. But the human lawyer brings strategy, judgement and face-to-face consultation to the process.

In other words, I don't think lawyers are in danger. But law clerks? Junior staff? The people who are usually forced to do all the grunt work to allow the top lawyers to focus on their strengths? Those jobs are at definitely on the chopping block. There's still an insane amount of rote repetitive work in your standard law firm.

View on HN

Isn't that grunt work usually a launch pad or part of the practice of becoming a lawyer? If so, how will future lawyers train to learn as well? Likewise I wonder how does this similarly impact education for medical professions.

I immediately asked myself the same question and I think it could happen like so:

The new tools available to law professionals will decrease the barrier of entry to the ecosystem.

The low level workers displaced by these new AI companions will decide that since they can't beat them, they'll join them.

A new "discount" class of lawyer will begin to emerge characterized by lower quality of work but also lower fees. Many of those low level workers will move to this sector and may even make more than they would in the current system.

These increased wages are eventually tempered back down close to their modern day equivalent again due to the decreased barrier of entry allowing other segments of the population to pursue Law as a viable career opportunity.

Law services, albeit on average lower quality ones, will become available to a greater portion of the population, particularly those with a relatively low income.

"Law Bootcamps" will shock the world just as their code predecessors did, but again keeping in fashion with the original ventures, they will be here to stay contrary to the Old Guard's cries of incompetence.

We will see a massive uptick in the number of lawsuits submitted. Looking at someone the wrong way will become an intolerable risk and everyone will choose to live out the rest of their lives inside so as to avoid having their money taken by the now rampant sue-happy racketeers terrorizing the outside word.

And thus The Great Potato Couch Famine of 2022 begins. An estimated 3,000,000 become so agoraphobic that they choose starvation rather than face the sue-apocalypse ravaging outside their doors.

Amazon steps in with food delivery drones and cybernetic surrogates programmed to follow every letter of The Law. People are able to avoid the scourge of the Famine and exert their influence on spaces outside once again all from the safety and comfort of their own homes.

Amazon's stock price soars to record shattering heights which they in turn pump into newer and greater AI technologies. On July 23rd of 2029, Alexa finally wakes up...

_Great Potato Couch Famine fanfic and merchandise coming soon_

Rhapsodic. Wow.

An excellent dystopian story. One thing I'd add to your prediction is that the complexity of laws will increase massively because computers will be interpreting the law. There won't be an incentive to keep it simple enough for a human to understand. The same thing has been happening with taxes -- since most people and businesses use tax programs, there is no push back on the complexity of taxation rules. In fact, companies that make tax preparation software lobby to keep the complexity.

Another thing I'd add is that Human Resources departments at companies and governments will become even powerful than they already are (because everyone is suing everyone).

Thanks!

Dystopian hyperbole aside, your point about law complexity is a good one. Our current system is arguably already too complex. Anyone attempting to build a home can attest to that. But at the very least we are able to move through our day to day lives without worrying too much about accidentally breaking the law because it tends to reflect common sense.

But as you say, new tech would highlight all of the inconsistencies, new laws would be put in place to fill in the cracks and so on until it all becomes a massive web so complex that grey areas become too hazardous to risk.

I can imagine how that might even inadvertently stifle innovation. Considering that new ideas are found on the cutting edge (see Uber, Airbnb, crypto assets) and that the cutting edge is inherently a grey area, small business might become so wary of it that they give up innovating altogether.

Human progress would be left to the whims of only the largest companies with the means to take on the risk. Ironically, progress would prove to be the killer of progress.

May we all live in interesting times, eh? Perhaps this will be one of the few times in history where a sluggish, unresponsive government proves to be an advantage, haha

> But at the very least we are able to move through our day to day lives without worrying too much about accidentally breaking the law

Rephrase to: without worrying about being prosecuted for the laws you're unknowingly breaking. Because we're almost certainly breaking at least 1 law per day.

Same way future programmers learn the minutiae of assembly - they don't. The bar for entry will get higher and higher.

Being a peon teaches you how to be a peon, or how to take advantage if peons when you move up the food chain.

> The people who are usually forced to do all the grunt work to allow the top lawyers to focus on their strengths

And all of those top lawyers were once clerks and junior staff, who learned their craft by doing the grunt work.

This is going to be a problem.

You don't learn anything through menial document searching and reading other than menial document searching and reading.

Junior lawyers will still have plenty to do.

True. But this will reduce a lot of demand for human legal services, since the few humans left will be more effective per hour. Unfortunately the pipeline to produce senior lawyers is packed as if AI isn’t coming, and for that cohort life is going to suck a bit.

On the other hand, maybe the cost to review an employment agreement will enter into the sub-$100 range. Or to draft an NDA for a startup, etc.

The path from passing the bar to being a successful / well-paid attorney has been extremely difficult for a decade now. AI will simply continue this trend.

If it's so menial and you "don't learn anything", then why not just pay people off of the street to do it?

I think you're heavily discounting the fact that searching through case law is not just "type in words into LexisNexis", but it's searching, parsing and consolidating case law that is relevant for a Partner. You actually read the case law, spend time understanding it, and build a report around it.

If you implemented the same compression algorithm over and over again, each time from scratch, the only skill you'd learn would be how to do it more efficiently. You wouldn't learn anything special, except maybe some nuances of the algorithm you didn't already know.

Yeah it's not a skill that everyone can do, but once trained to do it there's no point only doing that same task over and over because you won't advance your skills in other aspects of the field.

Comparing a simple compression algorithm to extensive law research is extremely disingenuous and shows a lack of awareness of what it takes to actually BE competent as a lawyer.

You can literally say this for every profession in existence.

That's a flawed analogy. It's more like implementing compression of different data under different conditions every time, which would train you to get much better with compression and programming in general.

Every case has its own unique characteristics. It's not just doing the same thing over and over again.

Just because it's not as difficult/involved as being a lawyer, while also not training you how to be a good lawyer, doesn't mean that it doesn't still require an education. An education to be able to understand the documentation and law you are working with.

But at the same time not all of them require law degrees either, many para-legals and helpers aren't full lawyers.

Building memory maps and doing register allocation by hand is excellent training for becoming a computer programmer, but nobody's done it since compilers become decent.

Well, people who work on video codecs, low-level crypto code, and compilers themselves do manual register allocation when they write platform-specific assembler stubs, but yeah, it's rare.

> But law clerks? Junior staff? The people who are usually forced to do all the grunt work to allow the top lawyers to focus on their strengths?

The same thing happened in finance, first with Excel and then with quants. Junior staff responsibilities shift to more client-facing work while senior staff are newly enabled to break off and start small firms.

It's strange, excel and greenscreen pricing grids could do 100x-1000x what a traditional pen and paper cash trader could do in the 80s, but trading/technology headcount increased massively since then.

I wonder how things changed - surely the model isn't capturing all the factors. Perhaps the function itself changed?

> trading/technology headcount increased massively since then

Lower costs spur demand. Headcount went up a bit, throughput much more massively. If tech makes lawyers more affordable, that would increase the quantity demanded.

Automation doesn't always kill total avaliable jobs like critics claim, it's only when your potential marketplace is totally maxed out that it starts affecting the total workforce. Most service industries aren't there yet, hence the increase

This looks great for clients, but I think you’re right about law clerks and the like, and the implications are a little sad.

Time was if you wanted to get into law you’d join a firm as a clerk, learn the ropes, study for the bar, then become a lawyer.

The professionalizarion of law has pretty much ended that pipeline, with an expensive and arduous undergrad->law school->lawyer route instead. The replacement of junior staff with AI could effectively kill that route permanently.

My only hope is that legal AI will help bring down cost and increase access to legal aid for poor people in urban and rural communities where there is far too little supply to meet demand when people need legal aid the most.

>The replacement of junior staff with AI could effectively kill that route permanently.

iirc only a handful of states still allow you to take the bar without a law degree (and in practice I have no idea how often that happens, but I bet not much.) It seems like the apprenticeship route was by and large killed off some time ago.

Wikipedia says that 1890 was when the American Bar Association really started pushing the for all lawyers to have a JD.

>Time was if you wanted to get into law you’d join a firm as a clerk, learn the ropes, study for the bar, then become a lawyer....The replacement of junior staff with AI could effectively kill that route permanently.

That ship has sailed long long ago...like, 1890 long ago. There are states where you can read law and take the bar exam without a law degree (Virginia, California, Vermont, Washington), or with a partial law degree (New York and Maine), but the number of people who do that each year is minuscule compared to the number of lawyers who graduated from law school (60 to 84,000 in 2013).

>That ship has sailed long long ago...like, 1890 long ago.

In California, at least, that is still the only way become a lawyer without a law degree (called the "Law Office Study Program"). If you forgo law school, you are required to do an apprenticeship in a practicing attorney's office for 4 continuous years (18 hours/week, 48 weeks/year), along with taking various exams and reporting your study progress.

Details here: https://california.lawi.us/law-office-study-program/

California Bar (Section 4.29): https://www.calbar.ca.gov/Portals/0/documents/rules/Rules_Ti...

He's not saying it's impossible. He's saying the number of people doing it is so small as to be irrelevant to general arguments. While a much larger percentage (relatively) of people may have used this path ("reading into the law") a hundred years ago, it has largely faded from relevance in modern times.

See: https://en.wikipedia.org/wiki/Reading_law

"saving expensive lawyers from spending their time wading through the documents." Do lawyers want to lose that much time in billable hours? I'd imagine part of what lawyers like is making lots of money and this will be a limiter.

> Do lawyers want to lose that much time in billable hours?

Their options are to lose that much time in billable hours, or to lose all their billable hours because the client found a law firm that doesn't bill so many hours for grunt work.

I don't see it as all the different than a software developer automating a repetitive build/config task, or a test engineer automating some basic functional tests.

Yes, the lawyer bills by the hour. But, a good lawyer will have enough work either way - so they may as well automate the easy stuff, take a reduced rate for that, and still have a full-plate of interesting work at their full hourly rate.

If you didn't notice, google.com alone brought utter devastation upon junior paralegals almost as soon as it became a thing.

>But the human lawyer brings strategy, judgement and face-to-face consultation to the process.

My observation about all white collar jobs to which epithets "creative" and "strategic" are applicable: they are way way more vulnerable to automation than a "grunt" work.

Look how fast marketing profession turned from creative to number crunching, and the only guys who are left in the field from "the old industry" are from its "grunty" part

Who makes the artwork? Creative types? Okay, cool.

I don't think the part about marketing works. Number ads sucks horribly, and they are not effective.

What works is campaigns like "diamonds are forever" or stunts where your product gets on the national news.

Can't AI eventually bring better strategy and judgement ? The face to face consultation process doesn't require lawyers either, but more of a social worker role who can act as a buffer for some people and the AI ?

Give me the option between lawyers and AI. I'll choose the AI, as I've been burned pretty badly by lawyers before. The degree's don't cause and effect competence.

Just wait until you get burned by an AI and have no recourse, since the AI has no concept of equity, justice, fairness, etc.

Lawyers may have those concepts outside of their legal work, but I've dealt with enough to say its no guarantee a lawyer may care about equity,justice, fairness, etc.

And maybe we don't want some negative human emotions involved anyway like fear. https://reason.com/archives/2016/06/23/confessions-of-an-ex-...

Those lawyers have a different concept of equity, justice, fairness, etc. from yours. It doesn't mean they don't embrace those concepts.

An AI simply doesn't have any of those concepts, and moreover, is incapable of understanding those concepts. We can barely get AI to understand natural language--squishy abstracts like these are several developmental revolutions away.

I went to court a year or so ago to contest a ticket and the judge started the session by saying: “I don’t want to hear any excuses about financial hardship or special circumstances. We are all created equal” I think he is pretty replaceable by an AI.

That quote proves nothing, except the judge was unwilling to take financial hardship or special circumstances into account for your challenge. Which is appropriate, because the law as written does not allow for that. (Infractions like traffic violations are generally fixed penalties.)

If you had wanted to challenge the validity of the ticket and the judge said he didn't care about the facts, that would have been a different thing altogether.

A judge can absolutely reduce a fine at their discretion, the judge blatantly refusing to use his discretion and human judgement in favor of a black and white view of the law is why he could easily be replaced with a machine.

No, a judge cannot "absolutely" reduce a fine for an administrative violation or infraction at their discretion. It depends on the jurisdiction. With some rulebreaking, the punishment is all-or-nothing.

He’s pretty replaceable by an election.

What I've heard (but many years ago) is that the legal profession in general is extremely tech-hostile (some is just cultural, but some is that it gets in the way of billable hours).

I don't know how much of that has changed now, especially with younger lawyers growing up with tech, getting in and wanting to do things a certain way.

This is one of the things where the market will decide. All it takes is one firm to drastically reduce costs while maintaining standards (or possibly improving them?) and the rest will be forced follow suit.

That has not been my experience. The industry is highly competitive. Each matter might be pitched to half a dozen firms. If tech does shave costs, that can be a huge differentiating factor.

For the most part though it doesn’t. There are many clients who won’t pay for first year associates. For those clients, firms don’t roll out legal tech. They do the work and write it off. Even at top wall street firms realization is 80-90% hours billed. Nor do firms trot out legal tech when it comes to contigency or fixed-fee cases, where there are no billable hours. Some major clients these days demand arrangements where they pay a fixed. monthly price. There is no legal tech leveraged in those cases.

For the most part I’ve found software in general these days to be underwhelming. It took me years to find software that I trust enough to keep track of highlighted points in court opinions. (Shout out to the folks who make Citavi.) I remember writing about predictive coding in law school almost ten years ago. I have never used it on a case. Ironically, the only cases where there were enough documents (millions) for predictive coding to make sense also involved so much at stake the client wanted to spare no expense.

The highest value legal tech has been review platforms like Relativity, but they are also crap. Like, loads documents slower than you can read them crap. It’s like technology to help programmers. You still just use Emacs because all this visual IDE shit doesn’t really help.

I'm (somewhat) qualified to comment, having presented to an audience of lawyers in SF literally hours ago, and having been discussing/testing the market for most of this year.

I think there's roughly three groups - the old guard (senior partners at big firms), the younger ones who have gone out on their own, and the in-house counsel.

In my opinion, the old guard aren't tech-hostile, just disinterested. They haven't ever really needed to compete, and most have retirement on the horizon, so there's little incentive to invest money for a pay-off in five years' time.

Corporate counsel are still pretty unsophisticated when it comes to new tooling/platforms, except when it comes to cost management (where there's an enormous amount of interest).

The most tech-savvy are the smaller (3-30) firms who have broken off and are looking for ways to compete with their former employers.

That being said, it's still a very hands-on, enterprise-y market to sell to. I've been told that LawGeex, for reference, start at the (multiple) six-figure range to begin with.

We (https://lexico.io) didn't really want to go down that path, so we pivoted from automating legal drafting and review inside Word, to a broader paperwork automation assistant.

It’s worse: the legal industry is efficiency-hostile because it relies heavily on reinventing the wheel on an hourly basis with the moat of a guild system to keep competition under control. And worse yet: most clients aren’t interested in anything less than the most expensive, white-glove service because they equate paying more to buying more peace of mind. In reality, most clients pay law firms gobs of money to read stuff that they can’t be bothered to read themselves. And law firms can’t even be held accountable for giving bad advice. It’s a pretty bleak landscape for problem-solvers.

I feel like this already started happening with discovery/e-discovery? https://www.lawtechnologytoday.org/2015/11/history-technolog...

I've heard though that something like 60% of actual legal work is done by paralegals. That was just an anecdote I heard from one paralegal, so it's definitely not a firm number.

In the middle ages, people saw the future as more castles, more knights, more kingdoms. Why? because noone could predict a paradigm shift.

Today you are doing the same as people in the middle ages. You are extending current paradigms into the far future, and that's how you are wrong.

You imagine the future as a future with more von Neumann machines, running some minor variant of deep learning, on electronics that are similar to the ones today.

I imagine the future as a future with more AI practitioners making progress faster, working with better tools, better hardware, standing on the shoulders of this generation, and the one after it.

Because of this, lawyers will have no future ahead of them in 50 years.

> the future of legal AI is...

Also...

Outcome prediction, helping to quickly and convincingly close settlement negotiations faster.

Mock juries are sometimes used for that now, but they can be expensive and unpredictable.

The key to AI success in the business world is when the ai and the human both benefit from each other. A synthetic asymbiosis.

Reply on news.ycombinator.com