Hacker News new | past | comments | ask | show | jobs | submit login
AI will cause “structural collapse” of law firms by 2030 (legalfutures.co.uk)
207 points by bootload on Dec 20, 2015 | hide | past | favorite | 148 comments

I'm currently working on a startup which supplies automated legal document analysis for the legal industry. [1] So I definitely see the power for AI for transforming the legal industry, but also encourage caution in reading the report.

One thing to note is that a huge amount of legal work done today is actually quite possible to automate. It's a lot of reading through documents and finding the few clauses that actually matter to a contract or document, then focusing in on those. It's defining an overall strategy and then fleshing out (mostly templated) documents to implement that strategy.

Another factor in favor of legal AI is that legal documents have much more restricted grammar than standard English text, leading to less ambiguity. This makes them both easier to parse and to create automatically (you don't need to make the text sound "natural"). By 2030, AI will be able to easily summarize a full legal document and provide relevant references. (In fact, probably much sooner.)

That being said, it's important to note that this report isn't predicting the disappearance of lawyers (nor am I). Instead, it's the collapse of traditional law firms, which depend on a pyramid of partners doing high-level sales and strategy work being supported by associates and paralegals doing the grunt work. The grunt work is rapidly becoming automated, but I don't think AI will ever replace the client interactions and recommendations component. We've seen a lot of success with AI-lawyer pairing and I do think we'll see much more of it in the future. This definitely challenges the organizational structure of law firms. In the future, I think we'll see a lot more small and individual practices where clients work with a partner they trust who can get a lot more done thanks to AI doing all the grunt work.

[1] https://docnavapp.com/

The true value of a lawyer is their judgement, analysis ability, and their power of persuasion.

To replace that you need a true Strong AI. So I expect lawyers to be replace shortly after everyone else is replaced. Give or take a few years, since lawyers have some protectionist options, like banning AI from being a lawyer.

Weak AI will eat around the edges. Better spelling, grammar, and cite checking software could reduce a lot of billable time. But most associates spend less than 10% of their time doing that.

For big litigation, document review has been mostly outsourced to legal service companies. I've been a junior associate (the ones traditionally responsible for reviewing documents) for about 2 year. I think I've billed ~200 hours to that (out ~4,200 hours I've billed total). Client won't pay our rates. So we hire contractors for 25 dollars an hour. There is still room for disruption. Predictive coding will probably become the norm within 15 years. But it's only going to put the temp "doc reviewers" out of job.

Westlaw and Lexis already summarize cases, index them. ALJ and Mathew Bender sells practice guides. Firms subscribe to proofreading, document processing, and printing services. They hire paralegals and assistants to do the less taxing work.

It's also important to note that business model of firms only relies on associates billing hours because it's convenient. Clients don't pay 500 an hour for an associate to make sure commas are correct because it costs that much to provide. It's just an add on service to get the high level talent's analysis and judgement. Nickle and diming essentially. Associate costs are fairly small portion of current billing rates.

Most of the grunt work I do is bespoke. Stuff like comparing one patent to another. We have google patent and questal to search. But it takes a human to do the analysis. Even humans of average intelligence with an engineering degree aren't very good at it. If you can build an AI that does it for any patent you feed in, all of our jobs are fucked.

> Weak AI will eat around the edges. Better spelling, grammar, and cite checking software could reduce a lot of billable time. But most associates spend less than 10% of their time doing that.

You must live in a world where NLP software is very advanced. In mine, researchers found it so hard to either produce or understand human language, that most of them switched to playing statistical tricks on images.

> The true value of a lawyer is their judgement, analysis ability, and their power of persuasion.

I agree. We're not going to be seeing AI replacing lawyers any time soon.

That being said, I think you are downplaying the efficiency gains which come from AI and lawyers working together, especially when it comes to bespoke work.

For your example, diffing two patents is indeed something which still requires human intelligence. But there are ways of making the process of comparison more efficient by automatically correlating similar claims, providing references inline, etc.

This goes beyond the easy wins which come from "better spelling, grammar, and cite checking software" (which are themselves non-negligible).

Yeah, "this" as they say.

People tend to think about job-destruction from AI in terms of when AI can replace everyone doing that job, but the real negative (in terms of job growth) impact comes a long time before that.

If AI is good enough to let 1 person do the job 5 used to do, that's already a huge economic impact -- 80% reduction is huge.

>providing references inline

Me and a few people at were joking that we should leave our firm and write an addon for Chrome that adds cites when you copy and paste from google patent.

I would kill for that.

If you could do away with Chrome you could maybe use IE + OneNote?

Cannot say for sure in your case but OneNote is pretty good at linking to where you copied stuff from.

What exactly would it need to do?

Does MacOS still support multiple format variants of copy/paste? Sounds like a small add on, in that case?

In general I agree, however in specific the tools you refer to are not well exposed to the general public. The more tasks I can do with confidence and without a lawyer, the better.

It's a bit like wordpress. Making a decent website really doesn't take a programmer today. But for a long time the frameworks (like django) were making site builders lives easy without making the projects near-free for end users.

The first high quality lawyer replacement service I've used is clerky. I expect more and more startups in that space.

Systems that use machine learning to do document review actually already exist. E.g. https://kirasystems.com. They currently still need a lawyer in the loop to review and adjust the results, but eventually will be accurate enough to produce summaries that are more accurate than a human without any intervention.

Traditional law firms are starting to invest in software like this. While it can be used to reduce hours, it can also be used to sell more by increasing the value of the hours spent. For example, instead of doing a review of a subset of a clients documents, they can offer to review all the documents for the same, or even slightly higher price.

Yes, I'm familiar with Kira (they're in a similar space to us). The technology isn't really doing summarization yet though, it's more like extraction/TOC building.

I'm definitely biased, but I do think that's the direction that legal technology is going. Right now it's about making it easier and faster for lawyers to review documents: eventually it will be about them not even having to review documents at all.

I'm curious: this discussion is all about improving software that understands legal language. Are there any (long-term) efforts to improve legal language to make it easier to understand for software?

It's my understanding that this project is doing that: http://codex.stanford.edu/, trying to turn law into something like computer code. See also: http://logic.stanford.edu/complaw/complaw.html

There seems to be a bunch of projects. Lacking an AI to do the job let me extract what caught my eye and seems to be answering the question.

>The Hammurabi Project is an experiment in converting legal source material to the C# programming language. [...] The idea is to grow a body of source code that mirrors the structure of U.S. federal and state legislation, regulation, and case law. https://mpoulshock.github.io/hammurabi/

>The Stanford Computable Contracts Initiative (SCCI) is working on legal technology that will help move the world from natural language based contracts toward a world of computable contacts.

> The project [Designing and Understanding Forensic Bayesian Networks with Arguments and Scenarios] ’s new approach is to link the successful statistical modelling technique of Bayesian networks to models that effectively dovetail legal argumentation and scenario construction in the legal world. http://www.ai.rug.nl/~verheij/nwofs/

>Instead of automating legal work, I intend to focus on augmenting legal work. Specifically, my project [Computational Linguistics and Effective Legal Drafting] focuses on using advances in computational linguistics technology to help lawyers draft more precise and error-free legal documents like contracts, regulations and statutes.

I wonder if these systems would be imperfect (and they'd probably be - I have yet to see a perfect software) how soon people start exploiting them by putting language in the contract that autoscanner would miss but that would give the originator who knows where the language is some advantage.

You make a great point about the success of AI-lawyer pairing. We need to avoid the Lump of Labor Fallacy[1], in which automation only replaces lower-level work. Advanced tools will allow less experienced (but, let's not forget, intelligent and highly educated) lawyers to perform more advanced work at an earlier point in their careers.

[1] https://en.wikipedia.org/wiki/Lump_of_labour_fallacy

Well, you certainly have a "dog in this fight." =) It is a little self-serving to say "Yep, this is possible and we are the ones doing it." But since you have such a startup, you'd be crazy not say that!

But as I have stated elsewhere, document summarization and classification is not even in the ballpark of an AI lawyer. Excel didn't replace the Accountant. It's like a fancy Excel for clerks and paralegals (and Excel was a wildly successful product).

You sort of admit that but I think it is more than just the interactions with clients or making them "feel good." I don't think it really addresses what lawyering is or how positive legal outcomes can be achieved. This ma be a weak comparison but it's like we all know that the point of programming is not writing code and cleaning up memory; it's to create solutions to problems.

Also, I think the traditional structure of the firm serves many more purposes than just specifying a hierarchy for task assignment.

> It is a little self-serving to say "Yep, this is possible and we are the ones doing it."

Fair enough. :) Though I tried to be a little more nuanced than that.

To be clear, I don't think AI will replace lawyers any time soon. It's just making them more efficient. Our customers are layers, not people looking for legal services.

> Excel didn't replace the Accountant.

Great analogy. We don't think AI is going to replace lawyers any time soon, but it will make them more efficient (and potentially eliminate some of the lower-tier staff, like what technology has done to pure "bookkeepers.")

> I think it is more than just the interactions with clients or making them "feel good."

Sorry if I didn't make it clear, but I do think lawyers still provide a lot of value which AI can't easily replace. There's a reason we didn't have an AI write our equity plan.

> I think the traditional structure of the firm serves many more purposes than just specifying a hierarchy for task assignment.

That might be true, but remember that associates are expensive. If the budget for them has pressure applied on it and with AI you can get away with 1 associate instead of 5, that's going to change the legal track substantially.

Fascinating! I do wonder what happens to the traditional white-shoe partner track, where associates make their bones do sixteen hour days of grunt work (albeit with fantastic compensation) as a method of gaining experience and reputation. The skills of a high-level attorney, regardless of the practice, are honed through decades of experience. How would systemic AI affect the development of the profession?

Don't worry, the AI isn't going the replace the "Perry Mason" kind of lawyers, they are irreplaceable! AI is designed to replace the "Junior Grunt Joe" kind of lawyer/associate who is mostly involved in documentation and assistance.

And where do the next generation of Perry Mason lawyers come from, if there are no rungs on the bottom of the ladder?

Well, that logic applies to all fields, not just legal. Lower rungs are disappearing from everywhere. In accountancy, junior accountants used to manually do "debit-credit" first, now Excel does it for them, so what will they learn?

In IT, junior programmers used to toil hard with algorithms and data-structures to come up with barely working programs. Now there are IDEs, libraries, frameworks, CMSs all built and ready for them, so what will they learn?

In Physics and Math, Scientists worked hard on just the theoretical framework first and scratch their heads in classrooms. Now there are computer software that do it for them. Perhaps this explains we didn't get another Newton/Einstein since last few decades?

Hence the article's thesis: Artificial intelligence will cause “structural collapse” of law firms by 2030.

Maybe they could actually help the poor?

We'll have to wait for AGI to replace litigators -- in other words, forever. :)

Not everything which associates do is grunt work, and certainly not all of it s going to be automated overnight.

Instead, what I think we'll see is that lawyers use AI to become more efficient, especially at the associate level. So the same workload which previously took 5 associates will take 1.

Firms will still be hiring associates and thus have a partner path, but intake will be much lower. So instead of having to be top half at a top 5 school you'll have to be in the top 10% to stand a chance.

The skills of a high level lawyer is in sales and negotiations and trials, not in the document analysis that junior lawyers do.

Just another tool to master.

If you've read Kasparov on playing chess with machines, he has some really interesting observations on practical and philosophical aspects of man vs. machine. His most interesting observation is that it's neither the best human grandmasters, nor the best chess machines, but rather the best algorithm for humans working with machines that tends to win. It does make for nice headlines when a machine like Deep Blue or Watson trounces the best human in a particular artificially-constrained test. But in the real world of unconstrained problem sets and complexity, my money is on the proper pairing of human and machine rather than one obtaining any lasting advantage over the other.

> Another factor in favor of legal AI is that legal documents have much more restricted grammar than standard English text, leading to less ambiguity.

That's...really not true. Legal documents pretty much admit the full scope of the English language -- both grammar and vocabulary -- though there are constructions and vocabulary that have special meaning in particular contexts within legal documents that would either not be found in English outside of the legal context or would have different meaning. (And ISTR seeing cases where construction within the particular documents led to significant questions of whether elements that exist both in general English and in the specialized language of law but have different meanings had their typical legal meaning or their more general meaning in the particular document.)

Can you recommend a reading list on the NLP / parsing / etc approach needed in this area - I recently had an argument along the lines of "of course you can breakdown a document in five lines of python" and was surprised to find it's harder than I thought (!)

>legal documents have much more restricted grammar than standard English text

What are the degrees of freedom that are enabled by grammar and differ?

Why isn't what you described possible with AI?

The high-level work?

Let me first say that I don't think we will have human-level AGI by 2030. If we somehow do, all bets are off.

The high-level work requires a large amount of creativity in developing nuanced strategies for a particular domain. It also involves integrating knowledge and ideas from a lot more than the law. Consider that the vast majority of intelligent humans who attempt to make it to partner don't make it. Top lawyers are highly skilled, and automating them away is not trivial.

Also consider the trust factor. We've done some customer discovery in this space and most people would just not be comfortable working directly with an AI. This is especially true because a good portion of what partners are doing is human contact work with clients, courts, and opposing counsel. There's a lot of leverage which comes from being able to read and manipulate people which AI can't replicate easily.

Note that I'm referring to the 2030 date specifically for this. I don't think lawyers (or, really, any current profession) will exist in 2100.

I don't see why "Uber for lawyers" wouldn't work.

I'm imagining a combined semi-automated service where the paperwork side of law - simple wills, notarisation, simple divorces, and so on - is completely automated, and the creative elements are farmed out to individual attorneys with a public track record of success who compete on price.

It's going to kill the traditional law firm business model stone dead.

Lawyers are basically paid to logic-chop precedent and written law, and to be persuasive and/or intimidating in more or less creative ways. IMO eventually all of that could be automated.

I don't see any reason why AI couldn't potentially have super-human powers of persuasion and manipulation. It won't happen by 2030, but the next step are document searching is to bundle up the basics of law as a low-cost automated service, and then start automating the process of argument, cross-examination, and persuasion - which will lead to an arms race, which will lead to very rapid advances, which will lead to complete automation of civil and criminal law by (I'd guess) the end of the century.

> I don't see why "Uber for lawyers" wouldn't work.

Primarily because people generally want to choose and talk to their attorney. It's basically the opposite of taxis, where people didn't care who the driver was and (generally) preferred minimal interaction.

For Uber, not having to interact with your driver or choose them is a benefit. In law, it would be a detriment.

That's why I think the future isn't a monolithic application where you interact with a faceless half-human, half-AI lawyer. It's small individual attorneys accomplishing a lot more without needing expensive support staff thanks to AI.

> complete automation of civil and criminal law by (I'd guess) the end of the century.

Okay, sure. In the longer term, every profession will be automated though.

My father insisted, back in the 90s, that people would all want their bills on paper. Therefore, he reasoned, the post office was safe for the foreseeable future.

I suspect your belief that people want to choose their lawyer might be mistaken because damn near zero people even want a lawyer in the first place. I just don't see it as the important factor, and you certainly didn't provide any support for your contention anyway. I suspect you're making the classic mistake of believing people want it because that's the way it is now, which is rarely actually true in my experience.

> Primarily because people generally want to choose and talk to their attorney.

IANAL. It's been a thing for very many years (since the late 1980s) because people freaking love talking about law, and they seek advice from random anons, (and offer advice to random anons) over the Internet.

There are plenty of people who seek legal advice from other people on Reddit.

A lot of people have very little money to pay for lawyers, but need legal advice, and so they're going to want anything they can get.

>Primarily because people generally want to choose and talk to their attorney

Yes but why do they want that? The reason is not just some emotional desire for hand holding. The main reason is that AIs are simply not (yet) very good at language related tasks, and there is not nearly as much progress as the current bout of hyperactive AI reporting suggests.

>I don't see why "Uber for lawyers" wouldn't work.

I'd imagine for the same reason HomeJoy didn't work. If you find a good lawyer - why bother going through the service and deal with the lawyer directly? I can see many firms adopting or licensing such a service, but I doubt it would change how clients approach them.

>I don't see any reason why AI couldn't potentially have super-human powers of persuasion and manipulation I'd imagine once we have this, we would have Strong AI. At that point many traditional business models would be stone dead.

Rocketlawyer, legalzoom, nolo , etc, are uber for lawyers.

Don't worry, I know you are being amiable. The current professions will largely exist in 2100. =) that's probably a safe bet. For "creative/intellectual" professions, a tremendous burden is on the possibility of an AGI Overlord to exist and for the "labor based professions," unless the "Post-Scarcity Utopia Event", there are currently people literally unable to afford an actual pot in which to urinate so I don't know if everyone in the future will be able to afford the Intel RX9000 Autonomous Plumbing Unit. =)

> Let me first say that I don't think we will have human-level AGI by 2030. If we somehow do, all bets are off.

If/when human-level AGI becomes reality, the economic impacts will be the least of anyone's concerns: https://en.wikipedia.org/wiki/Superintelligence:_Paths,_Dang...

Lawyer here. I do not know how soon AI will achieve either the humanity or know-how to replace lawyers, but will offer a few observations as a practitioner.

As for the know-how: Processing and comprehending legal texts does not seem to be an insurmountable technological task. However, the breadth of legal practice areas (family law, IP, criminal law, administrative law, etc.), and how these several areas often overlap to influence our daily lives seems a much more complex task.

As for the human element: legal problems are frequently extremely personal and require an empathetic response. This is so at the lawyer-client level, as well as the lawyer-lawyer level. After all, most legal issues require navigating complex relationships between humans, or groups of them. I don't think our society is ready for automated justice at this point. My clients certainly aren't.

So, it seems the human component isn't immediately threatened by AI, while the technical know-how might benefit those with deeper pockets and access to the "better" AI. Practically speaking though, the deep pocketed clients already have armies of lawyers at their disposal. With that in mind, the technical advances are likely to benefit smaller practitioners and clients (like me) who have limited resources.

I fully expect to be practicing law in 2030. If all the lawyer jobs are fully automated by then, I'll have MUCH more to worry about.

It's a common misconception that AI needs to be able to perform essentially any task a human can do to induce the "collapse" of an industry. But really, all it needs to do is make each lawyer 2x more efficient (by automating the easy-to-automate tasks) and suddenly half of all lawyers are unemployed even though every lawyer employed today does some tasks that can't be automated. That's a collapse.

This assumes there's a fixed demand for legal help. You might as well envision that the efficiency gains cause a huge drop in legal costs which unlocks a huge latent demand. For instance software engineers are probably 10x more efficient now compared to 20 years ago, but this does not mean mass unemployment.

Software is the exception since its a new industry that is simultaneously expanding demand by orders of magnitude while automation dramatically increases the productivity of individual SEs.

There are plenty of people who were in industry and were forced out because they could not adapt to more efficient tooling to keep their jobs. Those that learned on ASM and Fortran may not be adapt for a world of prebaked containers and building everything out of node plugins that do the work. Even less are they suited to training heuristics for genetic algorithms that implement some of our most sophisticated software projects like speech recognition or vision.

By comparison at least those with money have saturated their demand for lawyers. Law is a particularly targetable field since its often considered overhead or expense.

"Software is the exception." If I weren't reading this on HN...

Your only argument is stating that the demand for lawyers is saturated. How is that different from saying the supply curve and the demand curve cross? This is about a shift in the supply curve. The whole point of the previous comment is that AI-assisted legal counsel will be much much cheaper than it currently is. That means people will quite likely want more of it.

Software is the exception as-in it is the tool that is driving the elimination process itself. And at the same time it does lead to the same effects inside the software industry. These two things seem to be opposites but they aren't. They are different facets.

Akin to something that happened with accounting when computerised spreadsheets came about. Instead of killing the field it just expanded it as more people wanted advice and what-if scenarios explored.

And it's ludicrously obvious how much extra potential work there is for legal contractors. Right now there are companies turning down revenue-generating opportunities because the time/cost of scrutinising proposed contract changes isn't worth it. Not to mention a huge number of agreements done on a "standard contract, no exceptions" basis for the same reasons. If automation allows these to be turned around in hours, for a fraction of the price, the variation and complexity in standard contracts is going to multiply exponentially.

Even if the process of checking the contracts is fully automated, there's still a role for an associate/paralegal level lawyer (overseeing many more contracts than their predecessors) to answer followup questions, because their clients aren't going to navigate through case law summaries to get a more thorough explanation of non-standard term 3.1b's applicability to their specific situation if they can still phone somebody with a law degree.

Standard contracts exist for a variety of reasons, time/cost being one of them. Another one I'd like to highlight is the principal-agent problem[0]. If, due to automation, contracts went up by orders of magnitude in complexity then clients would be exposed to a new set of risks with a legal basis that they do not understand. The question then is: are these new, inscrutable risks acceptable to a client? Perhaps not, unless they are already a very large business.

[0]. https://en.wikipedia.org/wiki/Principal%E2%80%93agent_proble...

Do you understand that accountants are one of the first to go with the advent of AI? It's a foregone conclusion that accountants will not be around in 2030, at least according to Bill Gates

Judging by the hours my patent attorney friend works, automating 50% of what he does would bring him down to a 60 hour work week.

Well breadth of legal practice areas are what machines are good at, a machine that can analyze one law can analyze all of them and with constant quality. To give a concrete example, if some rather arcane part of insurance law may be relevant for a contract, a machine can then analyze the insurance law just as it did with contract law. By contrast a human lawyer has to ask for advice from a more specialized lawyer.

The human element is a quite good argument, at least without a major shift in perception of lawyers. However, I think that automation will realistically change what lawyers do, as well as the overall market for lawyers. One 'easy' area of law would be scanning employment contracts for standard clauses and you get some information at each paragraph what it means. Third page: first paragraph gets a green background and a note on the side says standard language, three more days off per year than standard. Second paragraph gets a yellow background and a note that says may get you into trouble if you use a company notebook for recreational programming. Third paragraph gets a red background and a text "Highly non standard, check with a lawyer." In this example, a lawyer still gets involved, he just does not have to check for the routine stuff and if history is any indication, this is what will happen. The routine stuff gets automated and the interesting cases are referred to the few remaining humans.

You're right that analyzing overlapping areas of law might also lend itself to an AI's strength. I guess I meant that applying those overlapping concepts to achieve a certain real world result seems to require input from someone with real world experience, and an empathetic side to fully understand the client's goals.

15 years is a crazy long time. Technological / scientific progress is exponential - we will witness much, much more change in the next 15 than we have in the last 15. And, that's a daunting thought, since 2000 was such a different world than it is now.

The only forces keeping law in its relative stasis are its attachments to slow-moving government processes, its 'guild-like culture' and the resulting protectionism. But, once capital moves faster in that direction (and it will) those systematic factors will slide apart quite easily.

It's good to be confident, but having participated in and thus seen the wave of progress in AI over the last 20 years, I'm pretty scared for my own career. It is definitively an exponential process and one that probably rivals Moore's law in effectiveness.

Most technology hasn't improved very much in 15 years. Technology developed in the 1980's and 1990's has found many new applications, because of the Internet, but fundamental systems haven't improved much. In the professions, the increases in efficiency from 1995 to 2000 were more significant than from 2000 to 2015. Even Moore's law has slowed dramatically in the last 15 years.

And the government and guild protections are fictions. I work at a small boutique that would love to be able to take bigger cases away from big white shoe firms by leveraging technology. We're not going to leave money on the table right now because of some abstract loyalty to the "guild." If the technology existed, we would use it. So would everyone else. The legal industry is extremely competitive. There are hundreds of large business law firms in the country. Way too many to keep up some artificial convention not go adopt new technology that works.

I don't think the government and guild protections are fictions. This article seems to be relevant:


I don't think they're fictions insofar as lawyers (and laws) command a unique place in our society that other industry inputs don't. Insofar as lawyers have the ability and the incentive to prevent sweeping changes to their industry, the guild mentality seems very real to me.

I take your point though: smaller firms are very willing to make technology changes to find themselves new competitive advantages. Here's hoping smaller firms like yours (and mine) keep nipping away at those monstrous firms =D. FWIW, I have not had positive experiences when dealing with them...

The article is about non-lawyer ownership of law firms. Loosening the rules on conflicts and non-lawyer ownership of law firms would enable what happened in the accounting industry: consolidation of the industry into a handful of mega-firms. I don't see how that would be conducive to innovation. Exactly the opposite is true. A legal "Big Four" would be far more able to resist technological changes than the hundreds of firms competing with each other today.

I'm trying to think of a concrete example of protectionism keeping out new technology, and frankly I'm at a loss. I don't think it squares with the economics. Clients these days come with a budget in mind. If a law firm can do the work with say half as many billable hours by using technology, they can bid to do the work for a fixed fee 25% below the market rate, and then pocket the difference. It'd be a no brainer, if that technology existed. Instead, more often you see firms agreeing to a budget cap and just eating hours billed over the cap.

No. Think back 15 years. Almost everything in technology was the same as today. Yes memory density is higher now, cpu power efficiency is better and display technology is somewhat improved but basically we're using the same technology now as then.

In my lifetime I'd need to go back to before 1974 or so for things to be really different. That's before microprocessors became commonly used. Before pong. Not before Unix though. 40 years. Still not that different from today.

Technically not much changed but the way we work and live is totally different because of the advances in tech. In 2000 internet was not everywhere; I was on holiday in LA in 2004 and the internet quality in hotels that advertised with internet was miserable and mostly unusable. Mobile devices sucked and we were still in the AI winter. Search engines in 2000 did deliver total drivel for programming related questions (and probably others).

Now my subscription allows me to use my 4g in most of the EU and the US, it works very well, wifi is everywhere so basically you do not have to remember any facts. Google AI recognises pictures and auto makes panoramas and videos from pictures. When I search anything programming the first hit is the answer. Children have smartphones and tablets and use them all the time. So do my parents and my 80+ old neighbours. There are millions of coders online instead of the tiny amount in 2000; starting a software services company in 2000 was a breeze (I did) to grow and get fortune 1000 clients as there simply was very little and most was crap; try that now. Startups are everywhere and a huge % of students see that as an option now. We can 3d print implants. Did you check games lately? I can go on and on.

Although tech only got faster and bigger the world is completely different. And that is because those 'marginal' enhancements in tech and science.

I'm thinking this must be an age thing. 15 years ago isn't a significant portion of my lifetime, so perhaps I see the progress in that time period as not being tremendous. Perhaps you're younger. I've been active in this business/space/industry since around 1975 and for me not a whole lot has changed. Sure today you can put a thing with the power of a VAX on your wrist, but we expected that to happen back in 1980. I moved into the house I live in now 15 years ago and immediately set out to build myself an internet connection. I used 802.11b products that are not too different to what we use today. The one thing that has changed in the past 15 years that I see clearly is there is much more data available on the Internet, but that's just a simple adoption curve effect. Back in 1992 we had the Microsoft Developer Network on CD with a search application that ran on Windows. It was basically the same thing as Stack Overflow, except of course much slower to update and featuring contributions from vastly fewer people. But essentially the same thing. I even remember having discussions in 1992 with colleagues about how the MSDN would surely evolve to be Internet-based rather than delivered by mail on a CD.

Maybe it depends where you are? East coast U.S. is entirely different culturally, behaviorally, commercially than it was 15 years ago. And, all that change has been enabled and propelled by a recursive process of capital and technology. Technology was barely a concept in most people's minds in 1999. Now, it is the prevailing culture.

I can't think of an industry that hasn't been touched since then, and most have been utterly transformed.

15 years ago I didn't think I'd see self-driving cars in my lifetime, and there didn't appear to be any prospect for computer vision to get anywhere near human levels of performance.

And these thing are still true.

We have self driving cars on the streets, and vision systems out performing humans at classifying objects.

No we don't. Where can I buy a self-driving car? The vision systems are performing on a statistical basis that is admittedly impressive, but they are not "outperforming humans at vision" in any way that I understand.

I totally agree but I think your defense is even too generous to the other side. For me, it is a little laughably naive, the lack of recognition of how complex "lawyering" actually is. Fancy NLP does not even get within several orders of magnitude of that.

Lawyers are meant to achieve positive results for clients even in litigious circumstances. For instance, in an auto accident case, a lawyer must carefully consider how best to present medical evidence to get the best possible offer in a settlement negotiation.

Lawyers must consider creative solutions to legal problems that are not as others suggested "just finding precedent snippets."

How do I even explain my complex case to the AI lawyer, assuming it is not some kind AGI Overlord? It's sort of stupid, I'm sorry to say, to me to see it suggested that lawyers can be replaced by LegalZoom+NLP. Maybe these people have never needed a lawyer for anything other than what they consider to be perfunctory paper pushing or they think we should not have an adversarial legal system.

If I have a serious legal problem I will take a serious firm of experienced lawyers over an NLP system. I guess if your outcome doesn't matter to you, you can pick what you like. It reminds me of the saying about the lawyer defending himself having a fool for a client.

> I don't think our society is ready for automated justice at this point. My clients certainly aren't.

They will if the service costs $10/mo instead of the $100s or $1000s per /hour/ you are charging...

I guess I should have explained more what I meant by "justice." In your quoted language, I intended it to mean justice as handed down by a person in a black robe.

Clients are already lapping up automated legal services (wills, contracts, etc.). But I don't think they're willing to accept a judgment from an opaque black box. Indeed, the cost to receive a judgment does tangentially involve some automatable processes, and those costs will come down as our computers get smarter.

To extrapolate a bit: as long as humans are determining legal outcomes, there will be human advocates. There will be human judges for as long as we have a Constitution.

They certainly won't if the cost of losing the case is sufficiently high.

Maybe the fact that I am not from somewhere where litigation pays off gives me a different perspective. Here suing is just not worth it usually as the judge will not demand significant payout. So when you win you will not get much anyway. Most of those cases can be replaced by computers today (judge and lawyers) and as far as I know those are most cases that are done here. So like trivial bankrupty cases, most small claims stuff. I had a few of those and I can and did just use Google to write my defense which works fine. No lawyers needed there. In the US the stakes are higher I guess so maybe I just cannot comment on that. And then there is criminal cases which might be probably more complex. Although AI might augment the lawyer to generate possible defenses and give the lawyer more creativity?

Interesting! What kind of legal processes do you already automate using software in your day to day life as a lawyer?

Our firm automates document assembly software for wills, trusts, purchase agreements, residential leases, and court forms.

We do some litigation, but in our small town it typically doesn't require extensive electronic discovery or document analysis that this post talks about.

As a side note, I'm a Freelance Web Developer. Below are links to my website and upwork profile. Let me know if you need to develop a web-based app or a mobile app around any legal concept and I'll gladly help!



It's worth reposting this 2011 NYT article, "Armies of Expensive Lawyers, Replaced by Cheaper Software"


> When five television studios became entangled in a Justice Department antitrust lawsuit against CBS, the cost was immense. As part of the obscure task of “discovery” — providing documents relevant to a lawsuit — the studios examined six million documents at a cost of more than $2.2 million, much of it to pay for a platoon of lawyers and paralegals who worked for months at high hourly rates.

> But that was in 1978. Now, thanks to advances in artificial intelligence, “e-discovery” software can analyze documents in a fraction of the time for a fraction of the cost. In January, for example, Blackstone Discovery of Palo Alto, Calif., helped analyze 1.5 million documents for less than $100,000.

> Some programs go beyond just finding documents with relevant terms at computer speeds. They can extract relevant concepts — like documents relevant to social protest in the Middle East — even in the absence of specific terms, and deduce patterns of behavior that would have eluded lawyers examining millions of documents.

The costs have remained largely static or even growing slightly because the volume of documents is increasing faster than automation is driving down costs.

That last paragraph is pretty much a description of google search is it not? That's what struck me about this article. It's describing AI as a thing which is coming in the future, but... knowledge mining document discovery tasks? The future is now! Google's AI (combined with the vast corpus of linked knowledge on the open web) is presumably already doing away with the need for a lot of low-level legal enquiries which would have previously been handled by law firms. I notice this article touches on the topic:


"Nonprofits and government entities have put almost all of the raw materials of American law online and Google and other search engines have made that law easier to find than ever. An American with a smart phone now has more access to legal sources than most lawyers or judges did fifteen years ago."

When stuff is made available on the open web for google (and other search engines) to index, that is the AI revolution right there. I say bring it on. Laws and court proceedings should be treated as public property. It should be put on the open web, permalinkable, annotatable, preferably with an explicit open license, so we can all benefit

What the article fails to discuss is what advances in the technology will push adoption over the tipping point. Are the algorithms going to get better? Are faster computers going to make until-now infeasible techniques practical? Is access to the Internet and big data going to make the difference? Legal automation technology has actually been around for quite awhile, and has seem limited uptake. So we're not at the point in the technology curve where we can just assume exponential improvements in the quality of the technology.[1]

And the limited uptake has not been for want of trying. There are plenty of firms that don't do document review in-house that would jump at the opportunity to take bigger cases away from mega-firms by leveraging technology. But the reality doesn't quite live up to the marketing. The overwhelming trend has been towards outsourcing work previously done by associates to low-priced contract attorneys, but even that shift seems to have reached a point of equilibrium.

[1] A good example is spelling and grammar correction. Word isn't much better at it than it was 15 years ago. You still can't do search and replace that fixes conjugation and numerical agreement.

As a person who is not part of the legal industry but had to deal with the legal process, I would welcome technology becoming more prevalent in the legal industry. My experience was especially tiring due to my residency and jurisdiction of legal case, I was part of, being in two different countries. So I would prepare the reply line by line running into pages during the weeks leading up to the hearing, take an overnight flight to the country of jurisdiction the day before hearing and all my lawyer did was turn up on the day of hearing with a printed copy of my reply wrapped around the standard legal header/footer and hand it to the judge. Everytime I turned up, I would find a new junior handling my case who didn't know half of the details since the firm takes on so many cases at once. Eventually I got tired and settled just to avoid the stress of spending weekends in airplanes.

If I had the choice of registering my appearance via a video conference link, it would have been so much more easier. If a never tiring bot would do the research, it could check for references/past cases which a junior can simply not check due to constraints of time/energy. My lawyer would have a much more exhaustive list of points to check and assess which applies to my case. So on and on.

Technology would result in law firms have to significantly re-evaluate what they do and I think it is high time that happens.

I'm not sure what country you are referring to but in some places you can video into the court room. I agree that's super convenient in comparison to what you are describing but it isn't exactly AI

Right about time. I had to deal with several immigration lawyers and it struck my mind how simple and small their decision tree is. A simple deviation from a standard case makes them want to "check and get back" which they probably do mostly by asking their colleagues.

Isn't that what you would expect? Most cases fit within a predictable mold, on others they consult with the rest of their firm (or case law, journals—whatever else it is lawyers consult).

Why would that be a bad thing?

False dichotomy territory there.

I had to deal with immigration lawyers for a friend about 2 years ago after deferred action took place. She was born in Mexico, but her parents came over here shortly after she was born. With deferred action, if you have a high school degree, are younger than 35, never left the US after you originally arrived here here and don't have anything on your "record" (which is a fuzzy thing in itself), you get a work permit with an SSN.

Now.. my friend is the most organized person you'll ever meet. For whatever reason, she decided to keep all of her homework, documents, mail, etc from high school 8 years prior. She never left the US, got mostly As, was good standing on jrotc. As a small anecdote, she also just got top employee award at a thousand person company without even working there for a year. Really, she was the kind of person that should have had a little bit of wiggle room.

Nope. She had a two month period of time that she couldn't find any documents from right after high school. They rejected her on the basis that she could have left the US at that time. It took 6 weeks to get a letter stating that. She immediately found some random mail and went through the whole process again. She got accepted, but it took another 6 weeks to get confirmation. THEN.. she had to 2 months to go to the SSN offices specially prepared for the "high traffic" where we were the only ones there, along with about 20 staff.

Oh, and then her brother who's been pulled over for speeding multiple times, dropped out of high school (he got his GED in fairness)... got a work permit with extreme ease. Hell, he even filed late without even using a lawyer.

Rant over, but you get the point. It's a frustrating process where there's a mold, but it's an ineffective baked-shit mold.

Edit: I helped another friend out who had a somewhat uncommon situation that let her get a greencard without going through the usual process. When DHS said she was good to go, the SSN told us it was "impossible to input the alien number" because it didn't have the correct amount of digits and couldn't be typed in their system. We were told to fill out a form that would take 6 weeks to process. DHS and the SSN office both refused to work with each other and we weren't allowed to use cell phones in the building. When we tried again about a month later, the person at the counter gave her an SSN on the spot without any issue. Frustrating.

Yeah, I would imagine that a large percentage of being a good lawyer is being able to recognize when a particular case is not a good candidate for routine handling. Same with doctors. On the surface, a case might be a good candidate for sending the patient home with aspirin (or whatever). Being able to recognize the unusual cases when that's not appropriate is where the real skill comes in.

Sometimes, of course, both doctors and lawyers screw up on this.

I wasn't saying it's a bad thing. It just seems like a job that should be fairly simple to replace once AI reached certain level. A lot of lawyers seem to be dealing with very few types of standard cases for which they repeat the same routine and there is no any 'problem solving' involved when you really need a high level of expertise.

That's like if you were a programmer and you had to develop the same kind of applications over and over again and when you could basically copy/paste most of your code from the previous projects.

It isn't, and I don't think shadow0 thinks it is either. The point I think he's trying to make is that making an AI to do their job for them would not be difficult

Yup, exactly this.

it's not bad per se, but the expense doesn't seem to be justified.

Imagine if an expert system did the work instead, at a fraction of the cost!

Should be less than a penny for many things, once all the patents expires.

Of course, we still may be paying a person to do a reality check on what a computer comes up with. The same way we will pay a doctor to confirm that, yes indeed, that leg has to come off.

Same with medical professionals

This is one industry that I would be happy to see go away. The concept that the amount of justice and legal protection one citizen can effectively receive depends on his/her ability to pay more than their opponent is very troublesome, and has resulted in many gross inequities in the United States.

And then the competition of AIs begins. Want a better AI? Pay more. Want the best AI? Get one that trumps all others. The principles are great, but in implementation, all is unknown, and I suspect more, not less, likely for manipulation.

That's why they started OpenAI I guess? To try to prevent that from happening. I agree with the parent that I would like that perverse and sick industry to die but I think that has more chance by replacing the judges with AI than the lawyers. Which won't happen in 15 years.

I know AI is a hot topic these days, and the technology hurdles are finally being overcome, but what about the human factor? It's human nature to want to control things, and integrating AI in a meaningful way into your life means giving up control of certain things. Delegating things to other humans is enormously challenging for many people (in fact, it's a key reason many startups fail!). I can't imagine delegating tasks with any sort of ambiguity to a machine being any easier.

You're right that the human factor is huge. I work in this space and from all our conversations it doesn't seem like any legal clients would be comfortable working directly with an AI directly. They value the human touch and advice provided by an actual lawyer.

That being said, there is already a division of labor in the legal industry between client-facing work and more rote document analysis/prep. The latter is where I think we'll see most of the automation: instead of being supported by a bunch of associates, the partners will simply work with an AI.

Client interactions and higher-level decisions will still be entirely human.

> Client interactions and higher-level decisions will still be entirely human.

I wonder if people ever said that about travel agencies?

I think there are plenty of things that people today wouldn't be comfortable with. I know many people that aren't comfortable with the idea of self-driving cars. 15 years from now, the things people are comfortable with could change.

"I wonder if people ever said that about travel agencies?"

Yep, they sure did. Or consider the restaurant business. With all other factors being equal, most of us would probably prefer a personal chef (or at least an expert one), an attentive, high-quality wait staff, fine china, real silverware, crystal glassware, linen tablecloths...

But they're not equal. In practice, McDonalds has a helluva lot more customers than Wolfgang Puck.

If this happens, it's probably going eat the legal industry from the bottom up. The rich will no doubt continue to have personal lawyers, just as they still have personal chefs. It will be the guys who want to contest a speeding ticket who will be the early adopters.

I agree that people will eventually change, but the timeline is more than 15 years.

Especially when you consider that the vast majority of legal customers are themselves more traditional. Comfort with pure AI for legal analysis will likely only come with the next generation of business managers.

> It's human nature to want to control things

I don't think that's justified. I don't even know what it means. Am I ceding control to a computer when I use software? I love using software to take things out of my control. A mind boggling number of things relevant to my life are out of my control and I have no desire to control them. I'm not particularly special in this regard: Most of the things out of my control are voluntarily out of the control of millions of others.

Human nature seems to be the rhetorical black box that justifies every claim about human beings.

I just think AI will be the first front people will use due to time and cost efficiency. If it doesn't seem that will work, then I'm sure the person would go to real lawyers after. That initial AI would change the legal game for many law suits though.

* Gets sued for bad product * AI finds 10 law suits in this persons past of the same nature that got nullified.

This would be fairly simple for the AI but would save a lot of time and money.

More than AI I would like to see standardisation and templates for common law problems. Like what creative commons did for copyright. Normal humans can use and very quickly understand them.

I have absolutely no interest in reading dozens of terms of service or purchase contracts I use every day, which are all wordy and hard to read, and say basically the same.

Man, I remember in the nineties when SGI declared they were going to solve protein folding. Still waiting on that one...

To be fair, NVIDIA was spawned out of several SGI engineers jumping ship as it was sinking, and protein folding is primarily done on GPUs these days.

A person could imagine SGI not collapsing and bringing GPGPU computing mainstream decades earlier.

Nearly half of my hours involve sitting down with clients to discuss their needs. I don't go anywhere near courtrooms. Having to draft a legal brief for a court, or even having to write a demand letter, means I have failed. I have yet to see any hint of AI capable of making such judgment calls on even simple matters.

How about this typical question: We shot a car commercial last month, but just noticed a McDonalds logo visible behind the car. The commercial is only being seen locally. Is it worth spending the money to pull the commercial and edit out the logo? (Spitting out trademark caselaw isn't an answer.)

Meanwhile, accounting firms look like their keeping their accountants, which seems like an easier task to automate away..

The keyword there being "seems". Most of the firms that I've worked with are actively automating or offshoring repetitive tasks. Much of the work that is being done requires meaningful critical thinking.

That's not really true. Back in the 1980s most small and medium businesses still used paper based journals. All the people who were employed to do that are gone - one person now can do the job of about 6 people in the 1980s.

Before the 1980s you had to do everything by post, not even fax. So you needed to follow up a lot more and talk a lot more on the phone.

Better technology doesn't necessarily mean that people will be "replaced" but it generally has meant that fewer people can do the same job.

Great article - I run a legal tech startup which automates some of the process work around litigation - https://rollrun.io and it's been receiving a lot of interest.

There have been a lot of the 'lawyers will be replaced by robots' type articles recently and I think the more positive spin is that lawyers will be able to spend their time more wisely with the help of AI. As a junior lawyer I spent a lot of time and effort doing tasks which were important to a case or a transaction but were more suited to a machine (the inspiration behind RollRun).

AI won't replace the skill and insight of a lawyer, but it will definitely speed us up and I'm grateful for that. And I'm sure clients will be too when it's reflected in fees.

At the top end, human expertise will still be necessary.

For example here is some recent work over the last few days at a leading London law firm.

" ... is advising GE on the intellectual property, know-how and information technology aspects of the potential sale of GE’s Equipment Finance and Receivables Finance businesses in France and Germany to Banque Fédérative du Crédit Mutuel.

We are working with Bredin Prat and Hengeler Mueller who advise on certain French and German IP and technology aspects of the transaction."


" ... advised Santander UK Group Holdings plc on its issue under its EUR 30 billion Euro Medium Term Note Programme of JPY 3,000,000,000 0.557 per cent. Notes due 2018 and JPY 27,000,000,000 0.787 per cent. Notes due 2020. The Notes are dual-listed and have been admitted to listing and trading on the Global Exchange Market of the Irish Stock Exchange and the TOKYO PRO-BOND Market of the Tokyo Stock Exchange, Inc.

Mizuho International plc and Nomura International plc acted as Joint Lead Managers on the issuances."


Bespoke work like this will still have human lawyers guiding and monitoring the course of the transaction.

At this level, the cost of elite lawyers is less than a rounding error on the balance sheet.

This misses that:

(1) Though the work in aggregate may be highly complex, it is disaggregatable. A very significant portion of the overall cost of these bespoke matters is composed of automable work (e.g., diligence, preparing closing documents). Humans may guide and monitor, but that could exist with significant changes to how legal work gets done.

(2) While "the cost of elite lawyers is less than a rounding error on the balance sheet”, most serious companies (e.g., GE) take efficiency very seriously in their businesses overall. Why would they not expect and push for it with their outside lawyers?

We need an economic system so everyone can participate in the benefit of AI and also significant economic incentives for AI to be optimized to do work for people rather than be self aware. Once this is in place there should be no reason not to push forward as fast as possible.

Funny thought on automated lawyers, to some extend lawyers are used to create standardized communication, aka contracts. So a law AI (lAI) would mean I give a natural language contract, my lAI translates to legalese and sends it to the lAI of the other guy which translates back from legalese to natural language. So legalese becomes just an interchange format for contracts, a badly designed protocol used only because we could not figure out how to transition to a better one. (Obvious JavaScript joke is obvious.)

I have a theory that about a third of all people's jobs exist simply to define a reason for having said job, lawyers even more so. The only ones incapable of seeing that AI will remove the need for many lawyers are lawyers who's job already is largely just them justifying the need for their services

AI may not replace lawyers themselves. But in my dealings with lawyers basic technology will mean that 1 lawyer can do the work of three.

For example: nothing is OCRed. So if a lawyer needs to get a section of a document they have to remember where it is.

There is some basic MS Word templating but really most law firms are stuck in the 19th century

There's likely still more than a little residual sting left from the era of WordPerfect 5.1, where there was a degree of sophistication (and sometimes enormous expense) that looked stable for a long time... then simply vanished. It's been 20 years, but there's something about the experience of having too much of your knowledge trapped inside machines you can't (practically) use anymore that's a little bit scary, and just doing the same thing with another word processor (that obsoletes its old files every now and then) isn't something most people would look forward to doing. Things will get better, but memories have to fade and assurances need to be built first.

This is sort of what's wrong with HN lately. There seems to be a sweeping majority of Singularity believers that just take "AGI in 30-50 years" as just an obvious inevitability. So therefore we just go straight into SciFi mode and make all kinds of speculation with that premise. Then, this couples with this sort of, I don't know how to explain it, maybe its like an uncritically overly anti-establishment or "naively disruptive" current that just collectively nods like "Oh yeah, the traditional law firm is dead meat."

It's surprising to m,e as I believe the community is largely made up of technologists, that I never see much questioning the fundamental claims here.

For that wet dream of that Internet Bayes Theorem Religion, that I will not name, to come true, it needs to first be possible and second, there must br credible, serious people to actually be working on it. Granted, artificial intelligence is a very hot area of research. However, for anyone that is either a researcher in this field, a graduate student in the field, or just someone actually very current on the latest papers and the current research programs, knows this is AGI of the SciFi sort is not really what is being worked on.

Not to go into all of the field but sure there are many interesting things happening in pattern recognition, NLP, machine learning and data mining and much work done with things like convolusional neural networks ..... but I'm sorry to say that these an AGI Overlord do not make. They are fancy techniques that yield very powerful and interesting results but in consideration of an AGI they are rather boring and banal.

Largely, the techniques have been discontinuous incremental improvements on reasonably simple ideas from the 1950s and 1960 but now utilizing massive parallelism. Unfortunately, I think, for AGI a non-trivial number of techniques are rooted in statistical and probabilistic methods, which I think anyone would agree is probably a a wrong track toward intuitive and creative problem solving.

It's not that those of us critical of AGI proclamations think that it is impossible, that would be rather bold, or that we think humans are magical or something. We probably just think that our brains, or even abstractly "true intelligence", are things that are unlikely able to be simulated by something like a Turing Machine. Any skeptic knows that extraordinary claims require extraordinary evidence. And I gotta tell ya, the way facial recognition works or even Watson, ain't it.

Also, I will say, just because something in the world is a certain way or if it is not exactly like Uber, does not mean it is necessarily a bad thing .... I'm just sayin' ....

An AGI Overlord would probably not be like the wild-eyed disruptor but more contemplative and understanding of the world as being very complex.

To follow on with your points, if the term "AI" had never been coined, and Terminator (and other such) movies had never been made, all the breathless articles would be about "learning algorithms iterated quickly, with simple inputs, simple outputs and a clear measure of success to help with the feedback".

Nobody would be under any illusion that this was anything we could label "intelligence".

Is there automated service that takes long end user agreements and makes them easily understandable?

There is Terms of service didnt read but that seem to be manual. https://tosdr.org/

I think Ai will take away the advantage larger firms have over smaller firms or individual lawyers. Large firms can and do bury small opposition firm or lawyers in grunt work that wastes time and people with few resources give up.

The phrase "structural collapse" does appear, to be fair, but what the report actually talks about is the hollowing out of low- to mid-level associate class that exists to do things like documents review and straightforward searches, not any reduction in the actual impact of law firms. Law firms will still exist, because (unsurprisingly!) they are legally mandated to exist. If you attempt to "disrupt" the legal profession as a non-lawyer you will be visited first by lawyers, and then by unfriendly men with guns. If you attempt to disrupt it seriously as a lawyer in a way that reduces its aggregate importance, you will be made illegal.

2030? I would say 2025 or even earlier, and not just Law Firms. If you have a white color job, AI is coming for it, and fast!

AI is just another abstraction, another tool to increase the power of the individual.

A simple lever allows one person to do the work of many, human technological development is all about making fancier levers (and people bitching about how the next one will end society)

I hope your joking or this sarcasm. If not, I think you should really read the latest journals and conference proceedings to see if you still think that ...

As others said already; you only need more efficiency, not full AGI to do that. And that simply is already happening.

They can say but that it still isn't true. Read the journals. Many of these professions which are supposed to disappear cannot actually be "made to disappear" because of a slightly better version of a primitive algorithm. It seems pretty obvious if you just stop and think about what would actually need to happen.

Autonomous vehicles, fine. If you ignore all the things that currently make transportation complicated, it's a fairly simple problem. I mean hell in the 1800s, they just put the vehicles on tracks with an engine and they were done. But most "white collar job" roles, if you get past cynical jokes, are actually well beyond the state of the art of AI research.

I mean, I love science fiction; I'm a big fan. But I'm also like a realist ...

I do not think they will disappear any time soon but I believe you need less people per case. So maybe there will be more cases and thus still the same amount or more lawyers required but one lawyer being able to do more cases assisted by some form of AI. And yes, most complete jobs can not (yet) be taken over but let's not forget the financials; if you can replace someone costing $40k per year by a less efficient service costing $10/mo then people will do it even if it does not fully replace the person. There are a lot of factors in that but for instance my accountant fired most of his staff when the SAAS package he uses got efficient enough; his right hand and him have to work slightly harder but he saves a very significant amount of money.

Similarly to how the industrial revolution 'disrupted' manual labor/blue collar jobs.

This article spells a doomsday scenario for the legal industry but having worked in a law firm during college I don't find that scenario to be very plausible.

Some legal work is just rote paperwork and perhaps much of this could be automated but it still requires a person to sign off on it etc. It remains to be seen whether this will just reduce the demand for lawyers. You are effectively just increasing the supply of legal services so while the quantity of services demanded (assuming some arbitrary quantifiable unit) will increase. It is not immediately clear that fewer lawyers will be in demand, I could see AI lowering the cost/barrier to entry for people seeking services and therefore more actual lawyers being required.

The rest of legal work is not just rote paper work and perhaps cannot be automated in the near future. If one person sues another person, I don't see any type of AI being capable of just calculating their way around who the problem and spitting out an answer for who gets what. Such a system that manages to account for every edge case would have to be unimaginably complex. You don't hire a lawyer to handle regular repeatable tasks that are well suited for computers, you hire them to handle things when something happens that is out of the ordinary or to consult on circumstances that are specific to you as an individual.

Law is also a field where reputation is tremendously important, and reputation isn't the type of problem that you can just engineer your way out of. Hundreds of years of tradition are not just going to be upended over night by some kid with a laptop because it isn't that sort of problem.

A prime target for AI is research. I could have saved a ton of time by scanning a document that gets OCR and then does a bunch of NLP to suggest related primary and secondary sources from a database. The specific algorithms aren't really that important but it would be a nice tool to augment the practice not replace it.

It's also important to note that the legal industry moves at a snails pace compared to the tech industry, people have a tendency to do things the way they've always been done and I don't think it is as prime for disruption as people on the outside from the tech industry would like to think. They will have a big impact to be sure but it isn't just going to crash like a tidal wave over night and put all these people out of jobs.

A further barrier is that all of this stuff still has to be adopted. Law is a very domain specific field that requires a lot of time and study, so is AI. The intersection of people who have experience in both is very slim indeed so the value proposition for lawyers at the top of their field who would be making these decisions is going to be a hard sell. It's going to be even harder when they realize it might decrease their billable hours.

tl;dr; AI will augment the legal field, not replace it. This viewpoint comes from my experience working in a law firm with software designed to automate some tasks, not pure speculation like this article seems to provide.

Lawyers won't allow it, I'm sure.


That's fine, as long as religion and faulty economics are also exposed as completely illogical by AI.

Religion, that's what you want AI researchers to spend resources on? Even if "exposed as completely illogical", it would not stop religion from helping people live their lives with comfort and guidance. I'm not saying everyone needs religion, but it does help many people. What reaction are you expecting the world to have in the scenario that AI shows religion to be illogical?

Religion has hurt a great many people too.

Religious arguments are full of logical fallacies. Getting those pointed out a non-human with no axe to grind might help a lot of people find there way out of this trap more quickly.

When I was 22, I remember a taped debate about religion I had borrowed from my brother. It helped me figure out that there really was no proof for all the assertions that I had accepted as true.

Being able to reach that conclusion without being in a debate myself was very useful. I wasn't trying to protect my self image or save someone else.

In much the same way, computer intelligence might be just the thing many need to help them out of the maze. Especially if they can have a conversation with it and it can precisely point out an error. And if the same intelligence can deal successfully with other complicated problems and negotiations.

It sounds like you are under the assumption that all religious people view their religion as scientific explanations. Religious fundamentalists take their religion literally, but not everyone is a religious fundamentalist.

Regardless, I think that "Religion has helped a lot of people, it's also hurt a great many people too.. so let's just focus on converting people out of religion!" is a dangerous position that will cause harm to many. Scientific processes for finding truth (which are really just things that are the least false) and religion are two separate domains. I don't see an explanation of what "the trap/maze" is and why it demands to be detached from humanity.

But you can do that as a human too. The problem is that all ends with 'but you cannot prove that He does not exist ans the bible says people will try to discredit Him'. Religious people are not all to rational usually so why even have the conversation? In the EU I see religion dying rapidly; I live in a tiny (40odd people with a bigger 100 people village in the valley next to it) village in the south of Spain where I moved down to from Amsterdam. In Amsterdam I was used to people being atheist. Even my maroccan friends who do go to the mosques only go because it is social thing not because they believe. Same for my Jewish friends. In Spain I thought the old men and women at least would believe and I was surprised to find that they actually almost all do not. 'Don't be silly, I go to church because after that we drink a lot' says my neigbour.

Not sure why the US is so religious but the education and lifestyle will kill it. No AI needed.

Not sure why the US is so religious but the education and lifestyle will kill it. No AI needed

I think you profoundly underestimate the power of both religion and philosophy on the soul and imagination of man. Back in the 16th century, the "new learning" confidently predicted it would sweep away religious belief and other forms of "antiquated" thinking. How silly and facile that seems now. The fedora-clad New Atheists are no different—arrogantly predicting the demise of something far more powerful and permanent than they are.

Like said besides the US and the middle east I do not see that in reality. And I do not underestimate the power of religion or philosophy: I just think only the latter will remain. Or hope it will. And it is not really silly yet: that is only a few 100 years ago. It takes time to get all humanity out of poverty, not dying young and educated. If we still believe in some intelligent entity(ies) that architected the universe and life in 2000 years you can call it silly. I for one find it silly that people believe in such an entity while I was raised and schooled as a christian until I was 20. But I started doubting it and asking questions that were answered with 'the bible says so ; do not doubt!' at a very young age. Which has the whole: everything points to this all being a nice story to keep the people quiet and docile and no science to back it up.

you can call it silly

But I didn't call it silly; I said it should be taken seriously. And frankly, "because science says so!" can be just as dogmatic (or even more so) than religion. "Science" isn't a monad anyway; we can't simply reference one overarching, completely non-contradictory science. And there are Christian approaches to science that, in fact, do not resort to "do not doubt." And that's just for starters—it's very misleading to posit a "religion or science: choose one" duality given the number of Christians in STEM professions, from Faraday to Larry Wall.

As far as "not sure why the US is so religious"—I don't think the US is more religious than other cultures, or why you would single the US out for approbation in that way. It's probably the case that Christianity has influenced Western civilisation and that has been expressed in a particular manner in the US on the Constitution with its enumerated rights and ordered liberties. The influence of Christianity on the West has much to do with why Western cultures are advanced in terms of technology, literacy, general learning, and the arts, even when those things are used for non-religious ends.

I do think the US is more religious than other developed nations. Take high GDP places like Scandinavia and Swiss and Netherlands which are atheist. That is why. And I agree with you about the influence. I do not understand how educated people think this can work and why the US with smart people works like that.

The USA isn't high GDP? Fascinating.

Ya, but you don't need high level AI to expose religion. A little first grade logic would do if people chose to use it in that manner. The fact that they often don't indicates AI won't change anything in this area either.

The true purpose of the legal profession, as with other jobs like accounting and IT, is to provide something to do for people to keep them from spending their free time challenging entrenched power structures. If bots can do some job in law, then the law will change to require more human audits of such work.

Because my comment follows guidelines, I presume some disagree with this opinion. Perhaps you're young and haven't worked out the true meaning of most paid work. I remember reading about a study about 15 years ago that assessed what proportion of various countries jobs actually contributed directly to production, e.g. mining, agriculture, etc, as opposed to "made-up" jobs like law, taxation, etc. Australia had a mere 7%, though other developed countries weren't much higher.

Not many jobs in law and justice contribute to catching and rehabilitating criminals who break laws that hurt others, and even then everything's done in triplicate (prosecution, defense, judge). Most legal jobs revolve around crimes that don't harm others, and civil and business law.

If the types of people who don't go into business don't have jobs, they might revolt and in some other way challenge entrenched power structures. So the government makes laws to maximize the number of people in work.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact