In today’s column, I am covering an extraordinarily popular topic in an extraordinarily unique way. My hint at the top here to whet your appetite is that sometimes Artificial Intelligence (AI) and especially modern-day generative AI can be instrumental in the most unapparent of ways. Generative AI, in case you don’t know, happens to be one of the latest and most exciting forms of AI and you might have used it or at least heard of it via the likes of apps such as ChatGPT, GPT-4, Bard, Claude, Gemini, etc.
The gist here is to bring to your attention the emerging importance of generative AI as applied to the law and as will inevitably be put into active and daily use by lawyers and judges, a topic that I have extensively covered in my column. See for example my analysis of AI in legal practices at the link here, AI for use by lawyers and law students at the link here, laws as applied to AI at the link here, lawyers messing up when wrongly using generative AI at the link here, and my popular fifty keystones that all lawyers need to know about AI for the law at the link here, just to name a few ways you can get quickly come up-to-speed on my multitude of AI-law analyses and predictions.
I will be relating this to something big that was in the news but that doesn’t seem on the surface to have anything whatsoever to do with generative AI and the law. You might be surprised by what I have to say.
A revealed mystery of sorts awaits you, hang in there.
Misconduct Hearing And Throes Of The Attorney-Client Privilege
Here’s the foundation for today’s discussion. I’ll get to the AI part of things once I’ve properly set the stage.
You would have to be living in a cave that utterly lacks any Internet service not to be at least vaguely aware of the Fani Willis alleged-misconduct court case or hearing that is taking place in Georgia. The story is being covered in banner headlines, streamed live on conventional news media, and beamed to thousands or perhaps purportedly millions watching on all manner of social media channels.
It is certainly one of the biggest courtroom real-life dramas so far this year.
Just about every legal and societal angle of the contentious matter has seemingly been heatedly discussed and debated. Plenty of talking heads have weighed in. You are welcome to do an Internet search to find out the basis for the hearing and why such monumental attention is being brought to bear. I am not going to take that already well-worn path of conventional coverage and shopworn analysis.
Nope, I instead want to bring to your attention something that few might have noticed. It is extremely subtle and virtually unnoticeable. Yet, in my view, it is a serious matter that has significant repercussions for how our judicial system works and how justice in the future can be substantively bolstered.
Let’s get into the weeds.
The hearing on the second day became bogged down in arcane legalese and many viewers might have wondered what the whole ruckus was about (if you mentally tuned out, that’s quite understandable). The courtroom was teaming with aggressively assertive lawyers on both sides of the case at hand. They were vigorously and, in a sense, desperately trying to argue their legal position in a determined effort to convince the judge of their rightful interpretation of the laws that might apply to the case.
Please know that this is not a jury trial and that the judge is considered the final trier of facts in this weighty case. This means that the lawyers must make their respective legal pitches to the judge. None of the usual cleverness and trickery that is aimed at a jury comes into play (you’ve seen those types of ploys in many movies and TV series about lawyering). The seasoned judge will be the one to examine the pleadings, review the evidence, and make a legal decision presumably based on the law as the judge interprets it.
Okay, getting back to the second day of the hearing, a crucial underpinning to the entire case entails whether a star witness who happens to be a lawyer can be legally compelled to testify about what they know regarding a former client. This clearly smacks of the revered attorney-client privilege, a topic hammered into our minds because of storylines in books and films, along with numerous real-life instances where the hefty consideration has materially shaped high-profile court cases.
Put on your thinking cap for the twist involved.
The client of the star witness is not only a former client but also a said-to-be friend. Whether they both equally perceive the friendship to still exist was openly questioned, but they freely separately acknowledged they both had been friends previously, during the time that the claimed legal representation occurred, including that they were also law firm partners around that same time frame.
I want you to pay close attention to the commingling here.
The star witness previously served as a lawyer who was representing his client in other legal matters, meanwhile, they also were law firm partners, and they were everyday best friends. Much of the time, an attorney-client privilege invocation is a lot cleaner since the attorney is professionally representing the client and there isn’t any other intertwining or murkiness about the relationship among them. You’ve got an attorney, and you’ve got a client. If the attorney was legally representing the client, the attorney-client privilege would usually be somewhat apparent (as I’ll be discussing momentarily, that lamentedly isn’t a complete truism either and wrinkles abound), and we wouldn’t be having angst about friendships and the like.
A conundrum arises in that (in non-legalese simple terms) where is the line between that which is information obtained under the protective cloak of the attorney-client privilege versus information that was garnered while having drinks in a bar and being besties talking about sports?
You are undoubtedly generally familiar with the classic attorney-client privilege precept, which it turns out rose to a tumultuous center stage at the misconduct hearing and legally burst on the boisterous second day. If the vaunted privilege were to prevail in this circumstance, seemingly the witness could not reveal legal-oriented facets about their former client who is also a former (or current) friend and former law firm partner.
Each side of the case has a different axe to grind as to whether they want the witness to be legally compelled to spill the beans. One side wants him to remain tightlipped and therefore are cheering on the protections afforded by the attorney-client privilege. The other side wants to open him up like a can of sardines and ergo stridently aims to pierce the claimed attorney-client privilege in this instance.
It is a legal standoff of epic proportions.
For those of you who watched the highly argumentative proceedings, you might have heard repeated references to the number 1.6. This might seem out of the blue to you. The attorneys were using shortcut language to refer to a rule promulgated by the American Bar Association.
Per the American Bar Association (ABA), there is Rule 1.6 entitled “Confidentiality Information – Communication” that addresses salient key points regarding the attorney-client privilege, such as this portion of the rule:
- “A fundamental principle in the client-lawyer relationship is that, in the absence of the client’s informed consent, the lawyer must not reveal information relating to the representation. See Rule 1.0(e) for the definition of informed consent. This contributes to the trust that is the hallmark of the client-lawyer relationship. The client is thereby encouraged to seek legal assistance and to communicate fully and frankly with the lawyer even as to embarrassing or legally damaging subject matter. The lawyer needs this information to represent the client effectively and, if necessary, to advise the client to refrain from wrongful conduct. Almost without exception, clients come to lawyers in order to determine their rights and what is, in the complex of laws and regulations, deemed to be legal and correct. Based upon experience, lawyers know that almost all clients follow the advice given, and the law is upheld” (ABA Rule 1.6, Subsection 2 excerpt).
I have extensively examined the attorney-client privilege in my column, such as the link here, and will simply note that this is a topic of a complexity far beyond what you might assume it is.
The nature of the privilege has a plethora of twists and turns, such as a famous decision handed down by the Supreme Court that sought to settle a matter known as “full and frank”. The U.S. Supreme Court case that was decided in 1981, expressed the vital nature of the attorney-client privilege and emphasized the full and frank nature:
- “The attorney-client privilege is the oldest of the privileges for confidential communications known to the common law. Its purpose is to encourage full and frank communication between attorneys and their clients and thereby promote broader public interests in the observance of law and administration of justice. The privilege recognizes that sound legal advice or advocacy serves public ends and that such advice or advocacy depends upon the lawyer’s being fully informed by the client” (Upjohn Co. v. United States, 449 U.S. 383, 1981).
You should also be aware that there are legally steeped exceptions underlying the attorney-client privilege. For example, consider this additional component of the formal ABA Rule 1.6:
- “Paragraph (b)(2) is a limited exception to the rule of confidentiality that permits the lawyer to reveal information to the extent necessary to enable affected persons or appropriate authorities to prevent the client from committing a crime or fraud, as defined in Rule 1.0(d), that is reasonably certain to result in substantial injury to the financial or property interests of another and in furtherance of which the client has used or is using the lawyer’s services. Such a serious abuse of the client-lawyer relationship by the client forfeits the protection of this Rule. The client can, of course, prevent such disclosure by refraining from the wrongful conduct” (ABA Rule 1.6, Subsection 7 excerpt).”
I am not going to get deeper into the attorney-client privilege topic since we could do so until the cows come home.
The emphasis that I want you to have as a takeaway is that the attorney-client privilege has a vast array of uncertainties and legal leeway. It is decidedly not cut and dry. You can spend a lifetime studying and arguing what constitutes the attorney-client privilege and whether it applies to a specific case in hand. I want to dispel any myths that there is somehow one crisp definition and that there is never any ambiguity or room for debate.
I would dare suggest that the open-ended facet of the attorney-client legal construct should not be a surprise when you quietly reflect on the torrid matter. I’m sure that you are viscerally cognizant that there is almost nothing about the law that is ever fully straightforward. Disagreement is always at the fore. Laws are squishy. Interpretations of what a given law means are always upon us.
In other words, there is always some different legal opinion or legal interpretation that enters the picture of things when you bring up the law. A cleaned-up version of an old line about lawyers is that when you see two lawyers and their lips are moving, you can pretty much bet that they are having a legal argument (the actual joke is much worse, but I didn’t want to go down that unseemly route).
One reason that we have so much difficulty with irrefutably pinning down laws is that they are said to be semantically ambiguous. I have covered this ambiguity in my columns because trying to apply AI to the law is confronted with once again the semantic ambiguities of the laws themselves. A hope by some is that we should just “AI program” the laws and turn the confounding ambiguity into crispness. The reason this is not straightforward is that laws are mushy, subject to alternative meanings, and cannot be quite so readily codified, see my analyses at the link here, the link here, and the link here.
I think this has made you ready for the big reveal that I want to bring up next.
Courtroom Pandemonium Of Smartphones And Laptops Abuzz
I will now take you into the courtroom on the second day of the hearing. If you have time to watch any of the recorded video of the case, you will relish doing so once I’ve aided in opening your eyes to what took place.
Are you ready?
Buckle up, here we go.
Even before the star witness got up onto the witness stand, a huge amount of rowdy wrangling took place amongst the many attorneys. There were almost a dozen attorneys on one side of the case and a handful on the other side. Do not envision the typical one attorney versus one attorney kind of situation. It was a morass of attorneys contending against a slightly smaller morass of attorneys.
They were all pitching the judge about what should happen once the star witness arrived at the court (curiously, it was stated that the witness had a doctor’s appointment that morning, even though the case was clearly scheduled, anyway he was arriving late, which I would say that any practicing attorneys watching must have found rather eyebrow-raising, and was indeed an intriguing curiosity; I’ll leave things there).
The reason I mention that the witness was arriving belatedly is that this allowed for lots of suddenly unoccupied time in a court brimming with lawyers. What to do? The judge tried to see if other matters could be inserted into the unexpected opening. Some matters were covered, but the unanticipated delay pretty much opened the door again to further haranguing about the attorney-client privilege that everyone knew was going to be invoked once the witness arrived and got into the witness stand.
In the computer field, we refer to this as the notion that whatever amount of computer memory you have, such as a set amount of disk space, inevitably will get filled up. If you unexpectedly find yourself in a courtroom that has open time, and there are eager and erstwhile attorneys there for a pending case, you can bet your bottom dollar they are going to try and fill up the time. There is eternal hope on their part that they can use their legal acumen to turn the tide in their favor. Might as well take as many swings of the bat as you can possibly get away with.
Anyway, the star witness arrived and with a flourish landed in the witness stand.
Game on.
Here we are now at the crux of my core premise.
Throughout the questioning of the witness, the attorney-client privilege issue continuously and vociferously interrupted the proceedings. Imagine this in your mind. Here’s the imagery I want you to envision. An attorney indignantly stands up. “You don’t have to answer that question since it encompasses the attorney-client privilege,” the attorney would exhort. A different attorney would leap up and bellow “The question does not fall into the attorney-client privilege and the witness has to answer the question.”
Rinse and repeat, dozens upon dozens of times.
The judge was logically being pummeled by legal claims on both sides. The testimony ground to a snail’s pace. Things got so bad that the moment an attorney started to ask a question, they were interrupted by some other attorney, doing so without even allowing the entire question to be asked. The judge tried to curtail the shenanigans at play, but the zany Wild West feverish clamoring was difficult to keep within the bounds of proper courtroom decorum.
As I said earlier, I am not getting into the whys and wherefores of what the legal issues associated with the claimed attorney-client privilege consisted of. If you are an attorney, you would certainly find the nuances and legal arguments of immense fascination. I instead want to look at the forest for the trees in this instance.
Allow me to identify the forest.
The attorneys were almost universally and fervently tapping away at their respective smartphones and laptops.
Why?
Because they were each frantically trying to bolster their legal argument by citing relevant case law. The judge was astute enough to immediately question any attorney asserting something about the law and probing what case law supported their posture. If you didn’t have your ducks in a row, the judge would pretty much chop down their legal argument into pieces or dutifully doubt the claimed position.
Rightfully so.
The judge can’t just let attorneys spout off about legal interpretations that lack proper and sufficient foundation. Sure, sometimes attorneys do so. They use a seat-of-the-pants approach. This judge wasn’t having any of that. You either knew what you were talking about, or you’d better sit down and be quiet.
Think before you speak was the unspoken watchword.
The attorneys realized that they would get immediately challenged, either by the attorneys on the other side or by the judge. Worse too, if you created a sour impression with the judge that you were speaking off the top of your head, this did not bode well in a case especially where the judge is the trier of fact. You would earnestly want to create an impression in the mind of the judge that you are making ironclad legal arguments. What you have to say is surefire. You are a walking and talking legal encyclopedia of irreproachable impeccably correct legal utterances.
Wait for a second, some of you might be thinking, wouldn’t the armies of attorneys in that courtroom for a case that is being globally live-streamed already completely know the law and thus have no qualms about speaking up?
Give me a moment to mention something that maybe requires a trigger warning.
Human attorneys are human.
Allow me to elaborate.
I would bet that most attorneys have found themselves in a similar predicament. No matter how much preparation you do for a case, once you are in the courtroom the real-time nature of the activities can shift or morph in directions you didn’t anticipate. The adage that once the first punch is thrown all game plans are set into disarray is common to wars, fistfights, and legal battles.
There’s something else to realize too.
For any non-lawyer layperson reading this discussion, you might be under the assumption that the judge would already know all laws and cases that apply to whatever the case in front of them entails. Sorry, no, that’s just not feasible. No human can have in their mind all possible permutations and combinations. The same could be said for attorneys. No matter how much they prepare beforehand, they cannot contain in their noggins all possible permutations and combinations of how the law is going to apply to the case at hand.
That might seem shocking. Aren’t attorneys supposed to know the law? Sure, they are likely to be versed in much of the cornerstone law pertaining to their specialty. They aren’t going to know all laws across all types of legal domains, nor know all laws and cases even within their chosen domain. No human attorney is going to know all the laws and all cases.
Not humanly possible.
The Information That Can Make Or Break Your Legal Posture
Assuming that a human attorney cannot know all laws and legal cases in their judicious minds, this takes us to what I refer to as the incompleteness principle. A savvy attorney must somehow arm themselves to overcome or contend with their incompleteness. You might be outgunned by the opposing side if you cannot muster the right information at the right moment in time.
In the past, lawyers would carry boxes of materials into the courtroom and be ready to dig into the heavy tombs of legal cases. The odds were that they would have earmarked certain passages and put hand-taped notes on crucial pages. If they were suddenly confronted with some aspect of the law that wasn’t at the forefront of their mind, they would politely ruffle through their materials. The opposing side would be doing likewise. Law librarians used to be highly prized, but nowadays they have been less so at the forefront, though a resurgence is said to be happening as a result of what I am about to mention. Hold on.
What does a lawyer in a courtroom in real-time do nowadays to cope with legal wrangling that is occurring on the fly?
The answer is perhaps obvious, they consult their smartphone or laptop.
This usually consists of bringing up various online legal repositories and doing hurried searches on the Internet. They enter keywords that they hope will snag the right cases or laws. The problem often is that you’ll get a zillion hits if the search terms are wide. You can’t take the time while the courtroom is rushing forward to figure out what’s golden and what’s junk. The other consternation is when you enter keywords that bring up no cases or no salient laws, or maybe just as badly bring up non-relevant stuff. You are wasting time and have nothing fruitful to show for it.
Let’s briefly talk about time.
I want you to be thinking of a sporting activity like say basketball. When the game clock is running, basketball players must be on the move. If you are aimlessly watching the fans and ignoring the game, the opposition is going to mercilessly score on you, as they should. Your mind must be in the game.
Furthermore, you have to make nearly split-second decisions. Should you pass the ball or take a shot? How long should you be holding the ball? The clock is ticking, ticking, ticking. Are you wasting time or are you using the clock wisely? All of this is going to greatly determine whether you will lose the game or win the game.
Shift to thinking about a courtroom when a case is underway.
In the misconduct hearing, the game clock was ticking, ticking, ticking. The judge had indicated that roughly the star witness would likely be on the stand for that one day (though, it was made clear that if needed, further time would be allotted if demonstrably required). The attorneys on both sides knew that they were amid a tersely timebound battle. Make your point now. If you missed the moment, you might have a devil of time trying to recover.
I realize that an attorney reading this would say that you can always submit something after the fact. Sure, that’s normally allowed. The problem is that you are then regrettably behind the eight ball. Why didn’t you bring up the matter during the hearing and in front of the judge with all the other attorneys present? Would the judge have done anything differently at the time of the hearing if you had done so?
A ton of second-guessing starts to permeate the attempts to deal with facets after the game is considered over or at least past the time of the basketball court presence.
You could almost smell the frenzy that was taking place in the courtroom. Keyboards were being hastily pounded away on. An attorney would stand up and bring a case to the attention of the judge. The judge would note the case and sometimes try to pull it up on their laptop or desktop. The opposing side would also rush to look up the case or law, aiming to figure out what they could do to shoot down the claimed indication.
My guess would be that attorneys watching the frenzy were merely nodding their heads in sympathy and empathy of doing the same thing, day after day. Same old, same old. Nothing new, though the legal cases themselves and the twists of the law about attorney-client privilege were abundantly engaging. The action of being frenetic and trying to search for your legally winning gotcha was old hat.
Those who are non-lawyers might have been slightly puzzled about what was going on. It just looked like one of those scenes you see of Wall Street traders wildly calling out arcane terminology. You probably didn’t observe that the attorneys were fishing around on their smartphones and laptops. You might not have noticed that they would read verbatim legalese stuff they had found. Frankly, it might merely have seemed like a circus that was without form or function.
I have now laid out a bit of a problem in the sense of how we today have attorneys and judges who in real-time are trying to feverishly carry on courtroom legal arguments when faced with timebound provisions, and that they often resort to looking up legally steeped materials on-the-fly via their handy-dandy smartphones and laptops or desktops.
Mull that over for a moment.
Sketchy.
I trust that you are saying to yourself, there must be a better way of doing things. This can’t be the best way to try a case. This can’t be the best way to resolve the incompleteness principle. Heavens, can’t we do better than this?
Maybe so.
Generative AI Coming To The Rescue
Let’s reflect on the challenges noted in the courtroom frenzy scenario.
One noteworthy issue that underlies the problem is that trying to explore online for relevant legal materials while seated in a courtroom and faced with a ticking clock amidst a topsy-turvy legally rapidly evolving environment is not easy. It is fraught with totally missing the mark. You enter whatever keywords come to your mind and then quickly inspect the results. Maybe, if you are lucky, you strike bingo.
Probably not.
Here’s how generative AI can come to the fore.
First, be aware that generative AI is a somewhat fluent-appearing capability that tends to exceed prior efforts at devising natural language processing (NLP). Via having been data trained across large swathes of the Internet, the mathematical and computational elements of generative AI have pattern-matched how humans write. You can enter prompts to tell generative AI what you want to do, providing the prompts in a natural way.
This is abundantly unlike the conventional NLP that you encounter in Siri and Alexa. If you’ve ever used Siri or Alexa, you know how exasperating the experience can be. You need to constrain your vocabulary, you need to make short utterances that hopefully won’t be grossly misinterpreted, and otherwise, you must interact in a stringent and stilted manner.
Consider two ways that generative AI can come into play when wanting to look up legal materials that might be on the Internet:
- (1) Generative AI with a web browser connection
- (2) Web browser that leverages generative AI
In the first instance, modern-day generative AI often has the capability to connect to the Internet and do searches on your behalf in real time. The contents being searched for can be based on a fluid conversation that you’ve engaged in with the generative AI. Likewise, the search results can be data examined by the generative AI and then elaborated or explained when presented to you. Thus, the point is that the generative AI is much more than just serving as a pass-thru. The AI can potentially spur the search and similarly spur how to relate the search findings to your matter at hand.
The makers of web browsers have realized that this capability endangers their seeming padlock on being able to browse the Internet. Why use a traditional browser if you can get so much more accomplished by leaning into generative AI? As a result, by and large, the browser makers have had to combine generative AI into their wares. Fight fire with fire, as they say. That’s the second means that I have listed above, namely, rather than directly invoking generative AI, instead do so indirectly via a browser that is connected with generative AI.
The same logic can be applied to accessing an online repository of legal materials. Whether you do so via the Internet or perhaps have some other means of garnering access, let’s agree that these two major methods apply:
- (a) Generative AI with an online connection to a legal repository or library
- (b) Online legal repository or library that leverages generative AI
I won’t repeat myself and will merely emphasize that once again we have either generative AI as essentially the front-end (point “a”) or we have generative AI as somewhat of the back-end (point “b”). Makers of online legal repositories realize that since generative AI is alluring to their legal clients, it makes sense to make their lives easier by integrally encompassing generative AI.
What does this mean for lawyers and judges in a harried setting trying to make or assess legal arguments?
They have a greater chance of finding the relevant laws and cases. They are likely to do so much faster and therefore be quicker to respond while the clock is ticking. They are less likely to waste time searching. They are less likely to stumble in the sense of finding something irrelevant and in their haste attempt to toss the found nugget into the legal discourse underway. All in all, they are bound to outgun their opposition, assuming that the opposition is not similarly armed. The side that is armed well will tend to prevail, all else being equal.
That sounds pretty good, which it is, but there are crucial caveats that we must acknowledge and consider.
Generic Generative AI Versus Legal Domain-Steeped Generative AI
Conventional modern-day generative AI is of a nature that I refer to as generic generative AI. The initial data training was done on a widespread basis across the Internet and involved smatterings of this or that along the way. Generative AI in that circumstance is not specialized in a specific domain and instead might be construed as a generalist.
If you want to use generic generative AI to advise you about financial issues, legal issues, medical issues, and the like and do so in the absence of a human domain expert involved, you ought to not consider doing so. There isn’t enough depth included in the generic generative AI to render the AI autonomously suitable for domains requiring specific expertise. The AI makers even mention this in their online licensing agreements when you sign up to use a generative AI app (most people don’t seem to read those agreements, surprisingly at times including lawyers, but they ought to be doing so, see my discussion at the link here and the link here).
Okay, so that’s the generic side of things.
AI researchers and AI developers realize that most of the contemporary generative AI is indeed generic and that people want generative AI to be deeper rather than solely shallow. Efforts are stridently being made to try and make generative AI that contains notable depth within various selected domains. One method to do this is called RAG (retrieval-augmented generation), which I’ve described in detail at the link here. Other methods are also being pursued and you can expect that we will soon see a slew of generative AI apps shaped around specific domains, see my predictions at the link here.
Here then is the rub.
If a lawyer or judge is trying to use generic generative AI while in the midst of courtroom legal argument melees, there is a chance that the AI won’t do much better than not using generative AI. The prompts you’ve entered might not contextually bolster the search. The prompts you’ve entered might not increase your odds of getting on-target laws and cases. And so on.
We need to then consider that if generic generative AI won’t necessarily improve your hasty efforts, will it somehow undercut them? In other words, we might be okay with generic generative AI even if it doesn’t improve matters. The rougher road is if the generic generative AI might make things worse than not having employed generative AI.
As they say, first do no harm.
I might also add that I don’t want to imply that using domain-steeped generative AI is going to be a nirvana. It won’t be. The odds of doing good are increased but it is not somehow an ironclad guarantee. You can still have facets that undercut the use of domain-steeped generative AI (I’ll mention those next)
Getting Snagged By Generative AI Gotchas
When using generative AI, you must be cognizant of the limitations and weaknesses. Period, end of story.
Sadly, as a prime example in the legal field, you might have heard about two lawyers in New York who opted to use generative AI for performing legal research and did so without suitably understanding what they were doing, see my detailed coverage at the link here. They included cited legal cases in formally filed briefs to the court that were cases fabricated or made up by the generative AI.
Bad idea.
Big mistake.
After the story went viral, the thought was that lawyers everywhere would henceforth not make the same mistake. They would use generative AI but cautiously double-check anything they obtained via the AI. This is a seemingly obvious and prudent step to take.
Not everyone seems to have gotten the message and we continue to have instances that keep arising. I’ve discussed whether it is incumbent upon judges to forewarn attorneys, which some believe is ridiculous that they would need to do so, see my analysis at the link here. There is also an ongoing debate about ensuring that law school students learn about generative AI at the get-go. The idea is that if they are taught about and utilize generative AI while earning their law degrees, they will be savvy about how to properly use generative AI in their legal practice, see my discussion at the link here.
Four key pitfalls of generative AI that I have extensively covered in my columns include:
- (i) Errors. Generative AI can produce wording and essays that have errors of fact or miscalculations, etc.
- (ii) Falsehoods. Generative AI can generate false assertions, which might be hard to discern as they are at times quite subtle falsehoods.
- (iii) Biases. Generative AI devises wording and essays that contain biases such as discriminatory indications.
- (iv) AI Hallucinations. Generative AI makes what appears to be factual content but is computationally made up or confabulated, and not factually based or grounded in reality (I don’t like the term “AI hallucinations” due to the anthropomorphizing of AI, but it seems to be a catchphrase that has regrettably gained acceptance, see my discussion at the link here).
Lest you shrug off those pitfalls and confidently believe that you would not possibly fall for them, please realize that people using generative AI are bound to land into the trap of accepting the interactions and outputs as stridently truthful and factual.
Doing so is easy-peasy.
You get lulled into assuming that everything generated is of the utmost correctness since much of the time it is. Maybe at the start of using generative AI, you are on your toes and double-check every little speck. After a while, you begin to let down your guard. You have inched your way into thinking that generative AI is flawless. This is very common.
Another factor that throws you off your Spidey-tingling sensibilities is that the interaction and essays generated have an air of confidence. This is by design, not by happenstance. The AI makers have devised generative AI to have an aura of authority. They could change this if they wanted to do so. Most people assume that the generative AI has this aura either by happenstance or is somehow a requirement associated with the underlying technology.
Nope, that’s false.
If you’d like to know more about the aura auspices and what might be done about this, see my analyses at the link here and the link here, for example.
Do Not Toss Out Generative AI On A Scare Or Whim
The kneejerk reaction to the noted limitations or downsides of generative AI is that maybe no one ought to come within twenty feet of generative AI. Stay clear away from it. A snap judgment would be that lawyers or judges are going to place themselves into a world of hurt by relying upon generative AI amid a turbulent courtroom legal argument battle.
This is the proverbial throw out the baby with the bathwater type of reasoning. It is unwise reasoning. It is lopsided reasoning.
When I chat with lawyers and judges, I point out that the very act of reasoning is part and parcel of what they do. They are legal reasoners. Reasoning is the crux of what they do. So, let’s use some reasoning here and see what we can come up with.
If you are going to use generative AI via your smartphone or laptop while in a courtroom legal brawl, you can readily use your legal noggin to gauge what kind of responses you are getting from the AI. You would hopefully or certainly be doing the same if you weren’t using generative AI. I dare say, and trust, that if a lawyer or judge did a conventional online search, they would still be scrutinizing whatever they find. They would be foolhardy to not do so.
Another emerging resolving aspect is the use of a second generative AI app that double-checks the response of the primary generative AI app that you are using, see my coverage at the link here. You would still want to keep your head in the game and do your own double-checking, but at least the chances are that some or a lot of the unsavory stuff is going to be caught for you. Numerous other methods of grounding generative AI to reality or truth are also being pursued, see my discussion at the link here.
Meanwhile, the generative AI on the balance is going to speed up your efforts by providing a better focus on what to look for, along with summarizing or explaining the potential reference. I would bet that someone in that setting who is armed with generative AI, and that knows what they are doing, can run circles around those that aren’t using generative AI in that same setting. They can peruse and assess more potentially on-target choices in the same time frame as someone being presented with a morass of possibilities.
I want to make sure we are all on the same page and that we are already agreeing that these rapid and determined attempts to find laws and cases are indeed already taking place, regardless of whether generative AI is in the picture. I say this because one retort or counterargument is often thrown at me that generative AI has no place in a courtroom because there isn’t any basis for using online capabilities while in a courtroom.
If that’s what you think, please watch the video of the misconduct hearing. Perhaps that will be an eye-opener.
I realize that a lot of the time, the use of online access is not undertaken during most everyday court cases. Some courts might even have restrictions or prohibitions about it. All in all, I would perhaps brazenly suggest that those days are numbered.
We are inextricably making our way towards online legal research activities while in the midst of a courtroom discourse, especially when you include the remote court activities that have arisen due to a follow-up to the days of COVID-19. I mention remote participation since then it is much easier to surreptitiously do online access to legal materials, doing so in a less obvious fashion via using your screen or maybe a second screen that you have set up in your home office. Nobody else necessarily realizes you are doing online searches, or maybe that your dog is doing them on your behalf (sorry, a bit of a joke there).
Deflating The Generative AI Wil Takeover Argument
I would like to address another false or certainly misleading quarrelsome positioning about generative AI in this legal application context.
It goes like this:
- Claimed position: The lawyers and judges in a courtroom that let’s say are opting to leverage generative AI will become non-thinkers. They will allow generative AI to do their legal thinking for them. Whatever the generative AI tells them, that’s what they will believe the laws or cases and the legal interpretation must be. In a sense, we are allowing generative AI to take over. The AI is autonomously going to decide the fate of those who enter our judicial system and are looking for justice.
Wow, that’s quite a leap, a Sharknado kind of jump-the-shark posturing.
First of all, you might want to take a look at my in-depth framework of the range of autonomous legal reasoning that we can expect AI to ultimately undertake, see “Identifying a Set of Autonomous Levels for AI-Based Computational Legal Reasoning”, Dr. Lance Eliot, December 2021, MIT Computational Law, at the link here. I have covered aspects of this topic of how AI will enter into the future of the law and society too in my Forbes column at the link here and the link here, in Bloomberg Law at the link here, the Law Society Gazette at the link here, on Law.com at the link here, and so on.
The crux is that we are not somehow on the verge of generative AI that is going to take over humankind. I realize that many banner headlines might lead you to think otherwise. Daily we seem to see clickbait indications of reaching sentient AI, referred to usually as Artificial General Intelligence (AGI), chillingly springing forth and we will be faced with an existential risk. The existential risk is that AGI is going to either wipe out humanity or maybe enslave us, see my scrutiny of these theories at the link here.
In a nutshell, I am asserting that we are not on the cusp of generative AI haphazardly and unexpectedly becoming judge and jury. This idea that lawyers and judges who use generative AI today are also going to lose their minds and cede control of the courtroom and legal pleadings to the AI is not simply farfetched but also dangerously alarmist.
The reason I claim this to be dangerously alarmist is that once again the counter-position seems to be that we ought to ban the use of generative AI to prevent this cataclysmic result. The ban needs to begin now. Do not allow a window of opportunity. If you do so, the AI will be like the camel that pokes its head into a tent, and the next thing you know, the whole camel is overfilling the tent.
I am of the belief that AI is going to gradually and incrementally be used in all phases of lawyering, including in the courtroom. This will be good. We will be doing this on a steady basis. It won’t be as though we have suddenly opened Pandora’s box and we are henceforth doomed. That is just wild talk.
I’d like to share with you one envisioned aspect that is technologically feasible even now, though we haven’t gone this route yet. I pointed out earlier in this discussion that the attorneys and the judge were looking up laws and cases via their online access. They did this by having to be cognizant of the existing legal context at hand, such as that the matter was dealing with the complexities of the attorney-client privilege.
Suppose that we had generative AI connected to the live stream or was otherwise overtly considered a tool by the court and had been engaged for use during the hearing. The generative AI could be scanning for example the audio in a mathematical and computationally pattern-matching way, trying to gauge what topics were at hand, doing so in real-time. It would have been computationally straightforward to detect that the attorney-client privilege topic was at the forefront of the courtroom discourse.
When the case seemingly got bogged down or went in a particular direction, it is quite possible for the generative AI to have been looking up suitable case law in anticipation of the courtroom getting into a legal wrangle over the matter. In that sense, generative AI could be proactive. Rather than solely having generative AI at hand to do reactive responses, the generative AI could be connected in real-time and computationally tracking what is taking place. The generative AI could then predict what might be a nearing legal conundrum and already have found pertinent legal cases and legal arguments underlying the underway topic.
A gentle electronic messaging nudge from the generative AI to the judge could be a signal to the judge that the AI has found pertinent materials. Of course, attorneys in the room would undoubtedly like a similar nudge. We will need to resolve whether this kind of generative AI use will be for all participants in the courtroom or just for some (this raises concerns that those armed with this will outdo those that don’t have it, putting some litigants at an unfair disadvantage; access to justice or A2J ethical issues abound). See my indications about these coming uses of generative AI at the link here and the link here.
I want to emphatically emphasize that this future of AI in the law will be in our hands. If we do stupid things when applying generative AI, we can likely expect stupid results. Enabling AI of any kind into the legal realm is of course going to need to be done sensibly, systematically, and with care. Doing so otherwise does carry especially significant risks.
Conclusion
I hope you found useful my observation that during a high-profile live-streamed court case we got to observe the use of online access as a means for forming legal arguments while the clock was ticking and as integral to a vitally important legal proceeding as it unfolded and evolved.
That’s the stuff of modern-day law.
Another element of modern-day law and the future of the law is the advent and adoption of generative AI. Better get used to it. Tilting at windmills is not going to stop this. The tidal wave is already underway. You cannot hold back the gates.
You assuredly can aid in guiding where the water is going to flow. You can assist in making sure that we are suitably going to use generative AI in the legal domain. The beauty and wonder of this is that we are currently at the start. What you do now is going to have a huge impact. How often do you get a chance to reshape an entire domain? I would venture to suggest not often.
Aristotle said this: “Law is order, and good law is good order.”
He likely didn’t know about generative AI, but we do, and we can leverage generative AI in a manner befitting his words of wisdom if we strive to do so and are mindful of doing so. Let’s get going.
