City Journal Winter 2016

Current Issue:

Winter 2016
Table of Contents
Tablet Editions
Click to visit City Journal California

Readers’ Comments

John O. McGinnis
Machines v. Lawyers « Back to Story

View Comments (35)

Add New Comment:

To send your message, please enter the words you see in the distorted image below, in order and separated by a space, and click "Submit." If you cannot read the words below, please click here to receive a new challenge.

Comments will appear online. Please do not submit comments containing advertising or obscene language. Comments containing certain content, such as URLs, may not appear online until they have been reviewed by a moderator.

Showing 35 Comment(s) Subscribe by RSS
I read the comments and find it funny that when one makes predictions about the future of anything then a chorus of nays pop up and scramble over each other to belittle the author's stance. Please, don't take yourselves so seriously---sit back and enjoy life as it rolls over you. Professor McGinnis is not the first to predict this particular development....In 1974 Lloyd Biggle Jr wrote a short novel (novelette) entitled "Monument" about a spaceship pilot who crash landed on an inhabited but unmapped planet. The author writes of the fight to save the planet from developers that want to take the planet and turn it into a planetary size resort. When the natives obtain "credits" (money) and hire an Attorney (Submaster Jarnes) the story shows the court battles on yet another world and those battles are fought with arguments and precedent setting cases that were fought in their past. The attorney merely feeds his case into the computer (even the Judge is a computer generated interface)and the computer weighs the two sides cases and makes a determination on the outcome. Submaster Jarnes and the planet "Langri" are far, far into the future and merely represent one path that the future of mankind may take. Lighten up Francis.
P.S. read the short novel "Monument". There is a short story by the same author and name but it, while good, is inferior to the novel and even progresses differently.

Great thoughts John! Speaking about disruption in the legal services sector:

My name is Jerry Bruckner. I am a New York Attorney for 20 years, and I have been extensively involved in developing and operating Ecommerce websites for over 15 years. I am the founder of, an online uncontested divorce forms preparation website. was recently launched. It is not a law firm. is an online self-directed legal document assembly service for people looking to file a simple uncontested divorce, where both spouses agree on all issues of their divorce, such as equitable distribution of their marital property and marital debts, spousal support (a/k/a alimony), child custody, visitation and child support.

For the low-cost and very affordable flat fee of only $95, customers can enter all of their divorce details into the website's easy to use online questionnaire (takes about 15-30 minutes depending upon the complexity of the divorce). Naturally, this service is not designed to handle very complex divorces... those divorces should be handled by very competent attorneys. However, there is a very large number of people who only need to file a simple uncontested divorce and this service is designed for them.

LetsGetDivorced instantly generate all of their required New York State uncontested divorce forms based on their answers, plus a custom marital settlement agreement ready to be signed and filed with the Court. Unlimited free customer support is provided via phone and email. Also, customers can make an unlimited number of edits to their information, and instantly generate and print out updated divorce forms.

While there are several similar companies, their websites are "clunky" at best, and many of them do not appear to be real companies.... just a website with no physical address or significant contact info listed.

For the initial launch, only handles uncontested divorces in New York State. The service will be expanded to additional states. At some point, I will reach out to the investor community to see if there is interest in funding the growth of this company.

I'd love to hear any feedback or criticism (positive or negative) about
I can be reached via

Commentary on the implications of tech on non-tech industries requires interdisciplinary expertise that almost necessarily eludes most who attempt it. McGinnis, familiar with the demands of the legal practice and a competent writer, clearly lacks the technical expertise to comment on the future of computing. He adopts, consequently, the approach of citing existing features without any explanation of their underlying mechanism when, really, it's only by discussing the latter that any predictive proposition can be persuasively advanced. Sure, chips are becoming exponentially more powerful and, sure, your phone now resembles a chess grandmaster, but I don't have sufficient technical knowledge to make anything out of these observations and, I suspect, neither does McGinnis.
In the Three Mile Island disaster the most important bit of information was not even imagined! That the emergency water coolant would FLOW UP HILL and out of the overflow pipe. It did so because steam pressure forced it out and the lawyers ops technicians believed that there was no problem because the emergency water coolant always registered FULL as it and the reactor emptied water uphill.

Why is this fact appearing in a utopian lawyer article? Well the law to quote someone resides in us all! And yet we remain ignorant of antigravity potentials in physical laws. So what of man-made-laws? There must be tons of loop holes in law. Do you want criminals to know how to escape retribution as lawyers and cops can?

My favorite case was a 'pound note forgery case' by an artist in UK that was an open & shut “GUILTY” case, but the 12 came back in less than ten minutes with not guilty. 2000 year old Latin that I've forgotten said motive!

Here in endth the lesson.

USA law has become statute based and plea bargaining and not ethereal. Everyone is guilty of some statute violation everyday. What makes the difference is who has the money to baffle the court with bullshti and who surcomes to temptation of lesser charges in plea bargains against their own innocence.

And the Judges are DUMB repeat DUMB. Oh I haven't offended enough legal minds yet the whole legal profession is corrupt, ignorant and selfserving. They do not see that the law is weakened by their criminal actions.

Case 1:
USA Torture for information Enough Said

Case 2:
USA Departments corruptly misinforming the citizen. Countless examples

Apple Monopoly eBook practices – No one looks at Amazon's tax or subsidies for selling eBooks at a loss for market share. The precedent for this case was the Japanese TV dumping trial in the 80s(?) where the DoJ dropped Sony because it sold the most expensive TV in USA. What does this say to your case?

What makes this case so anti your essay position is that there are hundreds of lawyers economists and interested onlookers who have different opinions. First the lawyers for the guilty publishers – what motivated them to make their clients plead guilty? Second why does Apple feel different? Third why does an ignorant in monopoly law (AS STATED BY THE JUDGE) court appointed inquisitor do demanding every instant expressions of remorse? Fourthly the stupid DoJ who find on guilt a gold mine to extract reparations for the ignorant main street buyers of eBooks and poorly managed states who find a winfall in money to fill their mismanaged coffers? Fifthly Economists who now join the bandwagon expressing opinions that Amazon's street price was subsidized and not a fair market price. And finally the publishers, again, who lost money to Apple's plan compared to Amazon's Who was the monopolist? What will algorithms say? The problem in maths as in life is that some solutions are impossible. Algorithms will show millions of Catch22's in law. The remedy is a Herculean task USA politics is incapable of solving. Hence my despair.

What you suggest in your runaway nuclear reaction of instant access to cases without due diligence is an acceleration of cases held eventually in-front of algorithmic Androids* ( Super Gods of DEEP THOUGHT might) without the human attributes of cause, mental state or motive. Innocents self incriminating at the scene on some Google Glasses device – Go strait to Gaol do not pass the court. And the rich Apparatchiks (yes Communist uber-citizens) free to criminal lives with impunity.

Another nail in USA's future.

* The metaphor is apt as Justice would be privatized or corporatized. Most probably Google would want in and thus a percentage of the fine. Miles better than being an ad placer. Its past CEO has spoken of no privacy. Now all the better to punish those thoughtless criminals and become rulers of the universe.
I heard a new A.I., when it pretends to be a small Ukrainian boy speaking English, can pass a Turing test 30% of the time...when it can pass the Patent Bar, I'll worry.
Just machines to make big decisions, programmed by fellows with compassion and vision. We'll be clean when their work is done, we'll be eternally free, yes, and eternally young...ooooh...
Ben, here's a bit of a lengthly article talking about the future of legal practice and how software, IT, machine logic will make inroads into the profession. You or Christin may already see it coming.
One of the clever things that people do in articles like this is slip in highly contentious claims as if they were fact, in the hope that the highlighted arguments will obscure the discrete controversial claim. It is clever but it is also very annoying when it leads to an entire argument being based on a false assumption:

>Lawyers in the modern regulatory state reap rewards from big government because their expertise is needed to understand and comply with (or exploit) complicated and ever-changing rules. In contrast, the entrepreneurs and innovators driving our computational revolution benefit more from a stable regulatory regime and limited government. As they replace lawyers in influence, they’re likely to shape a politics more friendly to markets and less so to regulation.

This is not a fact. It is an argument and a weak one at that. So why is it presented like one? Especially as, to my mind, it is entirely wrong.

The claim is predicated on the basis that lawyers only like regulations because they can use them to jack up their fees. This ignores the fact that regulations have inherent value. They aren't passed just to annoy companies or because of lobbying from lawyers. They are passed in an attempt to make things better for everyone. A regulation governing worker safety may be annoying for the factory owner who has to spend money complying with it but that is surely a better outcome than allowing unsafe working conditions leading to a worker being killed.

Further, it claims that markets operate best when lightly regulated. This is also not true. Leaving aside the obvious recent recession caused by poorly regulated financial markets, it is just not true that weak regulations benefit markets.

Look at web browsers. Would Chrome, Firefox, et al, exist if Microsoft had not been forced, by these pesky regulations, to stop abusing its dominant market position to keep them out? Of course not.

To frame lawyers as responsible for regulations for wholly selfish reasons and to claim that a reduction in the legal profession would lead to a reduction in regulations and thus a brighter and better world for all the oppressed business owners is a silly argument. It undermines any point that this article has.
While technology does help to deal with the incredible volume of documents created by modern litigation, it is also responsible for the problem.

These days large litigation can involve the review of hundreds of thousands of electronic documents. Machines can help with the load but ultimately the workload is continuing to increase.

However I think we are seeing a situation where small attorneys and firms find things difficult, while the large commercial firms continue to do well.
This story must have been written by someone with little to no actual working knowledge of technology or the practice of law, given the degree to which it overstates and woefully misstates so very much about both.
No, Nick G., they won't even be able to drive a cab. In 1995, the State Comptroller of Texas revealed that one-fifth of the attorneys in the state had failed to pay the license fee to renew their licenses. The State Bar, which is in charge of the licensing process, said, "It's not our job to ensure that lawyers pay the licensing fee. We're not a collection agency." Meanwhile, every taxi cab driver must be duly licensed to drive a taxi. Hence, in my state it is literally true that the license renewal standards for taxi drivers are better enforced than those for lawyers.
So not only will we have no lawyers, they won't even be able to drive a cab?
Thanks for this lovely article, Sarah Connor. You'd best be back to preparing for Judgment Day, though. The machines are coming.
Monroe S. David May 31, 2014 at 2:26 PM
A Lawyer uses the information to advocate for his client and to persuade another party or a Judge of his client’s just cause. Technology has changed the nature of the legal profession by freeing up the lawyer to do what he or she does best —advocate!

As long as humans disagree and do dumb things, there will be business for Lawyers. Its better than being an Undertaker; we get repeat business!

If you get thrown in Jail, call your I.T. person ... or maybe your laptop.
My God, all these presumably intelligent people commenting here and weighing in on what technology can and cannot do, yet none of them is evidently familiar with a slick little program called SPELL CHECK!

I have never seen so many typos and bad writing.
"As search progresses, then, machine intelligence not only will identify precedents; it will also guide a lawyer’s judgment about where, when, and how to cite them."

This sounds like a fine meta-analytical tool for law review articles, but it has virtually nothing to do with the efficacy of crafting legal argumentation in briefs actually considered by judges. Moreover, "assess[ing] the strength of precedent by considering the degree to which other cases and briefs rely on certain decisions" sounds like what any experienced lawyer does by merely scanning the Westlaw citator list for a given case or headnote.

In the vast majority of cases, in fact, judges don't make decisions about legal issues based on analyses of the weight of precedent. They mainly apply well-established law to the facts before them. While the occasional published decision will engage in comparative analysis of competing trends among courts to approach an issue not already decided by a binding authority, such treatments are exceptional, not typical -- and that's only regarding the published opinions, which are usually the most interesting. Not only do most written opinions, published or otherwise, implicate only routine citations of "black-letter" law, this output represents only the tip of the judicial decision-making output in litigation.

The vast majority of actual lawyering has very little to do with this kind of analysis, which in fact is itself a marginal aspect of preparing legal papers. Most courts employ a limit of 25 or, at the most, 40 pages for legal briefs. The action is in the supporting documentation -- the reams of affidavits with attached documents, i.e., the facts on which cases are mainly decided culled through the discovery process.

And while machine intelligence is a good substitute for much of the stupidity of old-style document review, it does not replace the process of investigating the factual grounds of legal claims or their defenses. This process is premised not only on document review but discovery tools such as depositions and, increasingly, forensic methods involving data but requiring human management and analysis. While we have all seen lawyers who take depositions as if they themselves were mere machines, the process of preparing for and taking depositions, and then translating the testimony elicited in that process into the stuff real litigation is made of. is not about to be mechanized any time soon.

The argument here is unfortunately classic ivory-tower stuff, mistaking the fun give-and-take of legal analysis in case books and appellate decisions with the practice of law.
How can a con law professor have a deep and nuanced understanding of either technology application or the realities of practice? He can't. But I can as a retired businessperson married to a Biglaw partner and father of a systems programmer.

The areas that can be automated by pure technology are important, but limited.
As my daughter point out, AI will continue to suck because it is no good at figuring out contexts. But the areas that can be rationalized by techniques businessmen and engineers have been using for many decades are immense. Lawyers have created a trap for themselves by making it illegal for businesspeople and process experts to be principals. So they are now businesses and flow processes that are managed by --lawyers. Rather amusing, actually to hear about this on a daily basis.

The fact is that as rationalization proceeds, the bespoke craftsperson becomes increasingly valuable. Their ranks will continue to expand and their income will keep exploding. Only some will be technologists. But they will be good at exercising judgment and dealing with complexity. Computers are useless for that and will be for many decades.

Meanwhile, Biglaw cannot replace the thin classes from the recession that are the only juniors with enough experience to be useful. In order to keep up PPP during the recession, it seems all the firms threw away the seed corn needed during the recovery.
I agree with David. Moreover, as a lawyer, Professor McGinnis is a failure at carrying any burden of proof with respect to his argument.

First, he relies on speculation; instead of citing existing and reasonably widespread technology, he points to "Watson" - a unique computer with a much simpler task that what he is discussing, responding to one question, then litigating any case. There is no other data he points to, just the one, unique, annecdote.

When we have cheap and commonly available "super watsons" able to not only answer one short question, but instead analayze thousands of pages of data, come up with all the relevant questions, and then search for the answers and differentiate between the many shades of gray off application of law to (contested) facts in case law... then maybe we will have something to talk about. But given the complexity of the thinking, and the need to understand factual contexts (e.g. what kind of conduct constitutes or does not constitute harassment, and how do you persuade this particular judge or jury), this computer will need have a higher degree of articial intelligence than the so called "singularity"; at that point, we might all already be welcoming our robot overlords.

Second, he doesn't understand the way technology impacts work. For instance, Journalism hasn't seen a decline in jobs. TRADITIONAL JOURNALISM has. Jobs have moved to blogs and other electronic media or nitch publications. In any event, much of that is driven not by technology itself but by how technology interacts with the wants of the consumer. Consumers want either "instant" news (the 24 hour cycle), or expertise and thoughtful analysis. The "middle ground" that traditional newspapers represented has shrunk, but other areas have grown.

E-discovery, for the moment, is vastly increasing attorney employment, because of the massive amount of data that needs to be culled, millions of pages today vs. hundreds or thousands in the past. Technology is already streamlining that with better OCR and pattern recognition, but is nowhere near the point of replacing human analysis to evaluate computer selections. Similarly, Westlaw and Lexis do reduce the amount of time needed to research any particular case. But that just means cases are better researched, billing is adjusted, and more time is spent on crafting better arguments or investigating and analyzing facts than was done in the past.

Templates and forms have been available for a long time, often published in reporters and state procedures manuals, so other than the internet making these things more available (and abusable) to the layman, technology has done little here. And to the extent that the forms are more available, this often works to create more litigation and work, rather than diminish it.

The real issue in legal employment is the oversupply of JDs and barred attorneys. That is due to the greed of academic institutions, subsidized by the government who defraud naive "pre-adults" into thinking that law school is a secure path to economic stability and success. Many law schools will close since this ponzi scheme (benefiting academics/administrators) has reached its tipping point, and the easy money is slowing down (see mortgage bubble). That has nothing to do with legal practice or litigation, generally.
" “Every practitioner knows that when a hard case arises, the law books are ransacked from the time of the Norman Conquest and the court blindly applies any absolute precedent that may have been found by diligent counsel.”

True in a perfect world where legal precedence rules the courts. In today's environment, from the lowest courts up thru the appeals courts, particularly on the federal level, and, of course up to and including the Supreme Court in D.C. itself, legal precedent may be shunted aside in favor of expediency and political correctness.

Show me any legal decision before the last quarter of the last century that held that a violent felon, engaged in a shootout with police, could be exonerated by the courts because he was acting in 'self defense', or that a government mandate forcing an individual to buy government health insurance was actually a 'tax', or that an individual could use the public bathrooms of either sex depending what what sex he/she/it felt they were that day.

At that point, a computer, regardless of how 'intelligent' it is, is out of its league.
Meanwhile newly qualified lawyers in London's top firms are paid record sums, nearly USD 120K.
@agwisreal "Question: what about the prospect that a sufficiently Watson-like machine can win almost any case on almost any evidence, simply by its superior command of the game and sport of arguing cases?"

The best anaology I heard was this: You like a guy or girl, and you want to move it to the next level, and you input all of your emails/texts/conversations so far into a program. The program then (correctly) advises you what to say/do over the coming weeks/months and then at the precise moment it says: Kiss him/her now. And when you do it, your risk of causing offence/discomfort to the "target" is less than 1%. When we have such a program that always works correctly 95%+ of the time, that's when lawyers should be worried.

I like the analogy because a very large part of law is dealing with people and will always be. One day we'll no doubt have a "Kiss-Now Program" but it's a long, long way away.

I'm a lawyer and I've seen conversations like this when people say "A-ha, you lawyer, you have a vested interest and are in denial." Please. I hated discovery and knew it was wasteful and inaccurate when I did it by eye (in the mid-90s). I'm all for self-help and better templates. Decline of lawyers has been predicted for 10+ years but I don't seee it yet. (Although I'm in-house.) What I do see is greater compexity in business, and a growing need to manage legal risk. ie. A strong need for legally-trained businesspeople, able to anayse problems, write properly, assess risk and options, and make ethical recommendations, be they practicing lawyers or not.
@agwisreal "Question: what about the prospect that a sufficiently Watson-like machine can win almost any case on almost any evidence, simply by its superior command of the game and sport of arguing cases?"

The best anaology I heard was this: You like a guy or girl, and you want to move it to the next level, and you input all of your emails/texts/conversations so far into a program. The program then (correctly) advises you what to say/do over the coming weeks/months and then at the precise moment it says: Kiss him/her now. And when you do it, your risk of causing offence/discomfort to the "target" is
"Their interest frequently lies in legal complexity and the uncertainty it brings."

More honestly and completely: the more irrationality and immorality it brings. The more complex law is the less meaning it has, making it ripe for amoral abuse.

"The decline of lawyers may therefore prove a boon to the rule of law and to market norms. Computational innovators benefit from capitalism’s process of creative destruction; their new applications transform industry after industry. Their success lies with a stable rule of law and relatively light regulation. True, once successful, innovators become incumbents and may seek to use government to hamstring new entrants. But the dynamism of technological acceleration will make it difficult even for big government to hold back waves of new “disruptions.”

One can only hope. How wonderful that would be. A maturing society w a properly infinitesimal lawyer fraction comprised of moral attorneys, attorneys w character. Heaven!
And yet the employment and incomes of lawyers has risen steadily over the last decade. I guess markets don't predict the future?
Chris Nicholson May 29, 2014 at 4:45 PM
Anyone interested in the intersection of algorithms with the law should check out Robot, Robot and Hwang.

One set of open-source deep-learning tools can be found at

They do text analysis very well, and very fast.
Paul MacDonnell May 29, 2014 at 2:31 PM
Wasn't that Deep Blue and not Big Blue?
See "Jarndyce vs. Jarndyce" and the "Office of Circumlocution."
I have so many issues w/ Professor McGinnis's argument that I could, if I wished, write an entire essay in response. I will, instead, confine myself to two observations: First, so much of what Professor McGinnis says is phrased in conclusory terms. Although he tosses in a number of anecdotes in an attempt to shore up his position, he fails to do the hard work of analyzing the claims he makes. As a law professor, he should know better. Second,it seems that the good professor is trying to carve out for himself a niche in "professor-dom", and that this article is a first step in that direction. Unfortunately, the decline in law school admissions will bring with it (as the professor himself points out) a corresponding decline in the number of professors and of entire law schools. Putting to one side whether what Professor McGinnis says about how much of a role computers will play in the practice of law, one thing is certain: law professors are now, and for some time have been, an overpaid luxury that students can no longer afford. If what Professor McGinnis predicts comes to pass (and I have very real doubts about the likelihood of computers playing as large a role in the day-to-day practice of law as the professor believes), then he should be quite concerned about his own future in his current occupation. But, based on my own experience, the dystopian legal future predicted by Professor McGinnis is still some ways off.
in some ways techno-economics is being proposed as 'the solution' to the legal problem.

people said this about westlaw and lexis when they began offering an alternative to manual search.

techno-econimics results in efficiency of profit creation by reducing labor variable cost via investment into a new technology (pre-overhead and overhead)

this does not solve any problems but makes them further entrenched. the problem if the legal system is NOT the symptom (an overbloated obese ratcheting guild that lobbies for political guarantees of more work) but the cause -----a political system organized around adversarial runaway legal systems, and the generql overgrowth of many beauracracies at the state and federal level. more than 1 of 3 american jobs (american workforce is smallest since ww2) is with a government agency if you include prison labor and full time welfare/disability recipients under 65yoa . a shocking but realistic statistic.

and many more are dependent on government contracts awarded to private employers.

the ussa and its bearacracies and adversial corporate legal systems are the underlying cause. you will not solve this problem by deploying technology to make it more 'efficient' . you will only succeed in kicking the political can farther down the road.

the group think of the legal professor class focussing on technology as economic progress is displacing the original thought capable of pointing this out.
As someone who worked with "big data" before it was called "big data", I can tell you that the benefits have been greatly exaggerated and the problems greatly minimized. "Moneyball" was more fiction than fact. It was a cute story, but the A's didn't win any championships and quickly faded back into mediocrity. Processing of data has become more efficient, but the problems of polluted data and confirmation bias have grown along with it.
Pretty much the same thing is happening in education. As a simple example, in one class I teach, there already are "no book"....everything is online.

But this is bogus education - having sources online is no guarantee that students will actually read the material. In fact, it's been my experience that forcing students to go online pretty much guarantees that they will not read or only quickly scan sources.

A similar thing happens in projects such as "report writing". Only exceptional students will download/read material from which they glean ideas to be used in their work. Most students download material, print it, staple it together and present that as their reports. They are totally befuddled when they're informed that's not what is meant by "writing" a report.

Having material available thanks to "machines" does not lessen the requirement for "hard work" that's always been necessary for mastering a skill or subject. The prevalent idea that hard work is no longer necessary because the material is "available" is a cruel joke on students brought up to believe that "hitting the books" belongs to a bygone era and is no longer crucial to their success.
I generally agree with the author but have one bone to pick. The author suggests that software programs are likely sometime in the foreseeable future to be able to pick out the relevant legal issues from the type of random collection of facts that the real world usually presents. As a law professor, I can assure you that skill is relatively rare and seems to be a large component of what separates elite lawyers from the rest of the pack. That's not to say that programs won't be able to do that with simple facts that have already been sorted for legal relevance (e.g., deciding that fault is an issue in the description of an automobile accident), or that someday the programs might be far more capable.
Deep Blue, not Big Blue. A computer search would have spotted that. But the day lies yet distant when a computer could have not only spotted a correction, but framed the story and strung together facts and logic to flesh it out.

Question: what about the prospect that a sufficiently Watson-like machine can win almost any case on almost any evidence, simply by its superior command of the game and sport of arguing cases? We've seen this done by top-notch human lawyers (think OJ).

Craig S. Maxwell May 28, 2014 at 3:13 PM
Er, uh.. "anything be more welcome.?"
Craig S. Maxwell May 28, 2014 at 3:11 PM
"A decline in lawyers’ social influence." That's akin to saying, "a decline in the influence of roaches and rats." Could there be more welcome?