by Dr. Lance B. Eliot
For a free podcast of this article, visit this link https://ai-law.libsyn.com/website or find our AI & Law podcast series on Spotify, iTunes, iHeartRadio, plus on other audio services. For the latest trends about AI & Law, visit our website www.ai-law.legal
Key briefing points about this article:
- Lawyers are continually having to try and predict what will happen in a legal case
- Legal Judgment Prediction (LJP) involves devising computer models for making such predictions
- The use of Artificial Intelligence (AI) is aiming to boost the predictive capabilities of LJP
- There are advantages and also downsides to using AI-powered LJP predictions
- If AI LJP becomes superbly predictive it could have a dramatic judicial impact
You might say that lawyers are inexorably in the continual midst of making predictions. They do so while in the middle of a legal battle, they do so once a case result is proclaimed and thus have to predict whether an appeal might follow, they make predictions before a case even gets underway, and so on.
Being in the legal field is a boatload of predicting, along with facing numerous underwater minefields and other hidden dangers when trying to make such predictions.
For example, when considering whether to take a new case, an astute attorney will ruminate on the nature of the legal issues involved and what the likely outcome of the case might be. A client seeking legal counsel will undoubtedly expect their potential legal advisor to offer some indication of how the case is going to turn out.
Typically, prudent lawyers are hesitant to make outright unequivocal predictions per se and instead proffer what they likely believe or generally anticipate the outcome will be, casting the prediction as shrouded in potential complications. Thus, somewhat safely distancing themselves from a seemingly ironclad prediction, emphasizing the inexorable uncertainties involved in making prophecies about any guarantee of legal results.
Newbie attorneys sometimes get themselves into quite hot water by making bold proclamations as though they possess an unerring legal-beagle crystal ball. Besides inevitably opening themselves to potential accusations or pursuit of legal malpractice, they are undoubtedly going to find themselves on the wrong side of a soured client that later on recalls how assuredly boastful the initial prediction was, particularly if the case goes astray of the anticipated outcome. Furthermore, a legal prediction without any semblance of qualification or leeway is at the margins of the expected ethical requirements for practicing law since it can unduly mislead a client.
In short, lawyers make predictions, but it is a touchy subject, fraught with difficulties and dangers, and entails reading tea leaves that might be messy and occluded, yet somehow has to be undertaken, preferably safely so.
Some attorneys feel like they never signed-up to be soothsayers and wonder how they got plopped into the business of being legal prediction seers.
It is rare that law schools focus much on predictions and predictive methods and therefore an attorney must via seat-of-the-pants craft and refine their own abilities to render legal predictions. It is one thing to read the daily news and pontificate idly about how the Supreme Court might rule, and something altogether more somber and serious when having to assess a case that you are personally taking on and having to predict the outcome, which will have direct consequences on you, your practice, your client, and the like.
Legal Judgment Prediction Is Afoot
Increasingly there have been efforts to utilize computer-based models to aid in making legal predictions. This field of study and application is often referred to as Legal Judgment Prediction (LJP).
For my in-depth research paper on this topic, see “Legal Judgment Prediction (LJP) Amid The Advent Of Autonomous AI Legal Reasoning” at this link here: https://orcid.org/0000-0003-3081-1819
The idea is to try and formalize the predictive process and supplement human conjecture with mathematical and analytic tools. Early versions of LJP consisted of spreadsheet models that were crude and simplistic. Gradually, advanced statistical models have been employed.
The advent of Artificial Intelligence (AI) as blended into the practice of law is further bolstering the predictive capabilities for attorneys. By using AI techniques such as Natural Language Processing and Machine Learning, it is becoming increasingly feasible to computationally assess a large corpus of legal cases and based upon detected patterns then make predictions for a newly presented legal case.
That being said, do not be misled into believing that today’s AI for Legal Judgment Prediction is somehow sentient or superhuman in being able to render case outcome predictions. That’s just not the case. In my latest research on the use of AI for Legal Judgment Prediction, I point out that we are merely still in what I refer to as the Level 2 stage, and have quite a distance to go before we can get to the autonomous futuristic kinds of AI capabilities that will arise once we enter into say Level 4 or Level 5.
Digging Deeper Into LJP
One subtle but important facet about Legal Judgment Prediction is the meaning of the word “judgment” within the moniker itself. Does the usage of the overloaded word “judgment” imply solely an outcome-oriented focus, such that it is akin to saying Legal Outcome Prediction, or does it mean any kind of judicial decision-making or potentiality?
Here’s why this semantic parsing is crucial, perhaps painstakingly so.
Legal prediction models and allied efforts tend to exclusively aim at the outcome or final judgment of legal cases. Though this certainly is warranted and inarguably bona fide as a focal point, I’ve pointed out that a legal case can fruitfully employ predictive powers throughout its entire lifecycle. What might happen, for example, upon initially entering a plea for a legal case? As another example, what is going to occur midway in a legal case as to how the next step of the case is going to come out?
There are many legal battles and skirmishes on the way towards a final ending of a legal war, as it were. Yes, you want to know whether the war is going to be won or lost, and in addition, for those in the midst of battle, it certainly can be insightful and instrumental to gauge whether each battle or skirmish is going to end up favorably or not.
As such, AI capabilities for Legal Judgment Prediction ought to incorporate stepwise elements into their predictive capacities. When immersed in a legal case, and at some specified stage in the lifecycle of the case, it would be immensely helpful to leverage an AI prediction LJP that could aid in assessing how the next nearest step is going to fare. Or indicate how two steps from now, or eight steps from now, what the success proclivity looks like. These indications could arm attorneys with useful insights along the arduous journey of trying a case, rather than only offering a one-time end-state indication.
This brings up another existing qualm about some aspects of Legal Judgment Prediction.
Even if we assume that LJP is referring to essentially legal outcome prediction, a begging question that can readily be asked is what the outcome consists of. A prediction that you will prevail in the outcome of a particularly thorny case could be completely upended by subsequent appeals. An attorney using an AI-enabled prediction system needs to know whether the “outcome” being forecasted is with respect to the (shall we say) ultimate outcome or perhaps some intermediary result that is going to ultimately be overturned.
Another concern is that the emerging AI systems for legal predictive purposes can oftentimes be inscrutable. The arcane mathematics embodied in a large-scale Artificial Neural Network, referred to as Deep Learning will rarely have any ready-made logical explanation to it. You might be delighted that the AI has predicted a favorable outcome for your case, but it does so without any kind of overt common-sense reasoning or ability to articulate how such a conclusion was derived. Whether you are willing to rely upon or trust something solely of enigmatic automation can be dicey and disquieting.
Finally, there is an enduring philosophical debate in the field of law about the intrinsic nature of the law and the role of prophecies and predictions. You might recall the famous words of renowned jurist Oliver Wendell Holmes, namely: “The prophecies of what the courts will do in fact, and nothing more pretentious, are what I mean by law.” One interpretation of this now-classic statement is that perhaps the mainstay of what lawyers do is prediction, more so than merely as an afterthought or add-on to the task of practicing law.
There is also the street version of legal predictive mechanizations. Some say that you can anticipate a ruling by a judge due to how they have previously made their rulings and by inspection of their long-running historical record of jurisprudence. That’s the scientific and abundantly rationalized version of LJP. Others suggest, perhaps partially in jest and partially in all seriousness, whatever the judge perchance had for breakfast on the morning of a ruling is more aptly to indicate how the judge will rule on a case in-hand (colloquially known as “digestive jurisprudence”).
No matter how one might view the law, it seems nearly inarguable that making predictions and seeking to make on-target predictions is inescapably part of being in the field of law. That is a fundamental truth and for which improving our abilities to make legal predictions, by means such as leveraging AI, can aid in performing needed legal efforts with greater aplomb.
I’m predicting an expanding future for using AI-based prediction in the field of AI & Law and you can assuredly hold me to that prophecy.
For the latest trends about AI & Law, visit our website www.ai-law.legal
Additional writings by Dr. Lance Eliot:
- For Dr. Eliot’s books, see: https://www.amazon.com/author/lanceeliot
- For his Forbes column, see: https://forbes.com/sites/lanceeliot/
- For his AI Trends column, see: www.aitrends.com/ai-insider/
- For his Medium column, see: https://lance-eliot.medium.com/
And to follow Dr. Lance Eliot (@LanceEliot) on Twitter use: https://twitter.com/LanceEliot
Copyright © 2020 Dr. Lance Eliot. All Rights Reserved.