I recently had the unpleasant task of working with some young attorneys on a contract. Lawyers tend to carry on like they spend their days splitting atoms, but reality is much different. They spend most of their days cutting and pasting from one document to another. Contract law is simply not that difficult, but it has been made vastly complicated by a massive bird’s nest of verbiage. Too many lawyers with too much time and too much software.
Of course, contract law should not be difficult at all. In fact, it is something that could easily be automated. After all, a contract defines a relationship. Party X agrees to do something and Party Y agrees to do something. Then there are consequences for when either party fails to do what they promised. Life is simply not that variable where there needs to be new language for each contract. There is a reason we have the expression “boiler plate” to describe contract language.
A system where the two parties could answer a series of questions about their expectations of the other in the agreement would be easy to code. Writing some code that would force compromise over contradictions would be a bit more complicated, but no more complex than answering an on-line survey. Once the parties agree, the system would get a signature from both parties and register the contract.
My bet is most people would find this much better than dealing with human lawyers, who tend to get competitive over trivialities. The machine would present both parties with the most common agreement for their situation and that would satisfy the parties in most cases. This sort of normalization would also minimize disputes as people accepted the rules, rather than look for ways around them.
Disputes would go before the machine. Rather than two lawyers trying to trick a human judge into going for their version of events, the parties would answer questions on a touch screen. If you think lie detectors work, maybe add those into the mix. Either way, the robot judge would calculate the odds of each side being truthful and then render the verdict most likely to be in the interest of justice.
Having been in court too many times, human judges always try to force a deal. It is part of their process. Imagine the parties answering the questions, given their odds of success and then offered a compromise. Since most contract cases are money cases, a haggling module would allow the parties to find a middle ground and that would be that. No deposition and interrogatories, just a couple of hours in front of iMagistrate on-line.
The beauty of the robot judge is he does not need to be in a courtroom. He could be a kiosk at the mall or an internet presence. Since the judge is automated, there would be no need for lawyers. Contract law would, for most situations, become a self-service issue, like pumping gas or washing your car. It would probably save Americans billions just in contract stuff.
Of course, there is no need to limit this approach to contracts. I am just starting there because it is fresh in my mind and one of the minor nuisances of my life. If you think it sounds preposterous, consider that you sign a contract every time you rent a car. The law regarding car rentals has become so formalized and regulated, it is just part of the background noise of vacation. Applying this to employment and services is no great leap.
Many criminal issues could also be automated as well. Drunk driving is just about to that point in America. The cop pulls you over and you are given a choice of taking a breath test or confessing. It has not put that way, but that is the reality in most states. If you refuse the breath test you lose your license until you see a judge, who always finds you guilty. Taking a breath test is just standing before the robot judge, when you think about it.
Imagine instead of arresting T’Quan and parking him in a booking cell, T’Quan is immediately brought before iMagistrate. The police enter their data and T’Quan answers a series of questions. Hard evidence like video and fingerprints are uploaded and the verdict is rendered. Maybe iMagistrate can offer the accused a plea deal. Either way, the process could take hours rather than months.
That will never happen, of course, because it is too frightening to us. Maybe not never, but no one reading this will live to see Robocop and RoboJudge, but the civil stuff is not unrealistic. There are on-line services for wills and setting up a business and even filing for divorce. Vast chunks of the law simply do not need a highly trained mouthpiece. They can be turned over to software. If software can write your will, it can sure as hell handle probate.
Got legal trouble? There is an app for that!
In the pre-internet days, Nolo Press used to put out some pretty good books with fill in the blank forms for the California market.
Speaking from 30+ years experience,much of this would be possible – but the acceptance curve would be extensive and lengthy. My practice is limited to workers’ compensation and related litigation. In my state, this is primarily an administrative practice involving heavy use of forms, much as suggested here. It can work and even contested claims can be handled quickly and efficiently, depending almost solely on the lawyers involved.
(contract law…)but it has been made vastly complicated by a massive bird’s nest of verbiage.
Well, you see, after the plagiarism of cut-and-paste from documents written by folks actually familiar
with the law (usually written by undergrads and “paras”), THEN comes the mandatory “mad Libs” fill-in-the-blank
bit to break up the pattern recognition.
It doesn’t go over so well when an “opinion” of law is so easily recognized as a “disclaimer” from “elsewhere”,
(with the “disclaimer” bit removed), nicely printed on expensive letter head/stationary, and charged US$ 2,000
fee…down the line.
We’re actually remarkably close on the contract side. Have PoC’d two packages that will do side by side contract review, will pull comparable language from prior contracts for comparison and has sufficient cognitive capabilities to recognize language regardless of format. From the perch of my particular industry, don’t see legal jobs as a growth industry.
After I posted this one, I remembered the name of the company that started the self-service legal business and checked their site. Legal Zoom has an amazing number of self-service options. For small business people, most legal work will be handled by software in the next decade. Leases, service agreements, commercial real estate sales, etc.
To some degree, the massive inflation of the regulatory code is just a jobs program for attorneys. There’s a limit to this.
We’re using it at the F500 finserv level. Lloyds now uses to do contract checks for “layered” insurance programs. But the cognitive capabilities allow the programs to learn and suggest language based on both prior art and judicial outcomes. Watson (which I’ve done some early stage work on) is a whole different level. What is more interesting is the degree to which the developers have really not thought out the downstream implications. Had a dinner table conversation with the IBMers after a demo of one of the early oncology treatment applications of Watson a few years ago, It was pretty spiffy, especially the capacity to ingest clinical trial data real time. So I asked how easy it would be to tune the economic algos to weigh patient age against cost so only the optimal economic outcome was recommended. “Pretty easy”. And patient/doctor would not see other recs? “nope”. So have you been working with an ethicist on this project? “Hmmm that would be a good idea”. Nuf sed.
funny that you mention this. i’m using nolo quicken willmaker plus for the first time to make a will and related boring documents. it’s almost fun, though, just answering a long checklist then it spits out a will. much cheaper and faster than a lawyer, and the concept is pretty close to what you are describing.
I wonder how many lawyers and law school students will soil their undies when they read that.
Sounds great, especially that robo-judge. If the state brings a case, the judge rubberstamps what they do. And, after getting billed $50 per email from my lawyer, for stuff I could have written, robo lawyers would be a lot cheaper.