AI In Law

Artificial intelligence is rapidly changing, the legal profession helping lawyers research cases, prepare arguments, and draft court documents. But a series of recent U.S. cases involving fabricated citations and inaccurate AI-generated material has exposed a serious risk. What happens when a lawyer relies on a tool that sounds convincing, but is simply wrong. On Taze Legal Perspective, we examine how AI is entering the courtroom, the accountability issues it raises, and where the legal system should draw the line. For an expert's take on the matter, we now connect with Professors Hong Zeyan at Kyeongi University. Good morning, Professors Hong. Good morning. Thank you for joining us from a crossy-Pacific ocean. It is. All right, so maybe we should get our listeners caught up first with the recent cases. Can you maybe walk us through the recent American cases that become warning signs about the potential misuse of AI in court filings? What exactly went wrong? Well, they are recent by definition, I guess, because the AI boom is in itself a very recent. The first case, the most famous one that I think you mentioned is that in 2023, two New York lawyers submitted their brief, infested with AI hallucinated citations, and the judge fined them for about $5,000 each. From there, there is a 2025 case in California, which the appellate court published the wrong citations cited by the attorneys. And then there are misspecific case that you specifically mentioned that there are four lawyers involved on both sides that submitted the AI infested wrong citations. And there were even the judges, federal judges, who pulled their opinions because they're law clerks, the court clerks used the AI and used a wrong citations. So it's not just lawyers, it goes up and down the profession at this point. I think that what's happened is, first of all, the convenience of using the technology and the allure of it is pretty strong without looking at the risk of hallucination and risk of being wrong about it. And then there is a economics of it. A lot of clients think, well, the AI is about, they want the legal product faster and cheaper. So I think there's this kind of a squeezing bench that is happening that's kind of blurring out and coming out in the court cases, sanctions, and some of the lawyers being disbarred. And it's not to dismiss artificial intelligence altogether. I use it almost every day, Professor Zong, and it's really helpful. It's like having an intern assistant help with the research, but of course, and then we do all the fact checking and we do the cross checking to ensure that because in the pursuit of mass communications, there's a vast amount of information available online that we can we through and fact check. But I mean, I guess time becomes a constraint, as you've said, there are economic interests that I guess both sides, both the client and the legal professionals might want to work out. Sometimes we forget that a lot of what lawyers do and what legal teams do is really not that sexy. Not like suits. It's a lot of paperwork and weeding through a lot of cases, too. But it is a warning sign nonetheless. I mean, should there be more rules and regulations that make the usage of AI crystal clear? All right. So let's talk about that penalty you just mentioned. When a legal professional files unverified AI-generated material, what consequences can follow? You mentioned dismaring. Well, that's one of the consequences. But there is actually an escalating level of consequences that lawyers face. First of all, if you submit a wrong or the faulty court documents, then it's to your disadvantage. So you could lose the case and your clients get upset about it. And then there is a next level, probably, the sanctions can be imposed by the judge, like we have seen. The New York case, $5,000, but we saw $15,000 sanction, $25,000 sanction, it could go up. And then there's a third level, which is you can get referred by the court to the state bar in the case of U.S. attorneys. And the judge, first of all, cannot disbar anybody, but the state bar, they can investigate and they can make a decision to disbar or sanction you from practicing law temporarily or permanently. So that's a serious, serious disadvantage for the profession. And then there's a, also, you can get sued by the client. Not only they get upset, but the malpractice of law suit is pretty serious one. So the professional's own financial future is in jeopardy as well. So there are levels monetarily as a profession and the monetary sanctions and even the license itself is at risk. I guess one more I would, especially if the New York lawyer and the appellate court decisions, your name as a lawyer is, you know, we'll go down the history as a bad example of what not to do. So I guess that's a death knell for your professional career. Can we talk about where the responsibility lies to if an AI system generates a false precedent or an incorrect interpretation of the law or creates, well, the side of materials out of thin air and that air ultimately contributes to a wrong verdict? Who should be held responsible? Is it the lawyer, the judge, the court or perhaps even the AI developer? Well, first and foremost, there's going to be lawyers. Those are already in place in a kind of a simple analogy. You can regard AI as a legal assistant who is not doing their job. When you cite a citation wrong or put in some wrong information in the court document and signs your name, you cannot go to the court and in front of the judge you say, "Hey, it's not my fault, it's my legal assistance," or "it's my law clerks. It's my computers." You cannot do that. So it's pretty clear that the lawyer has to be responsible for whatever he submits. I think that even in the AI's case, to a point where the AI completely replaces an attorney, I think that the same rule could be still workable. When it comes to the judge, if the AI comes into making a judgment, probably it's institutional in nature, so it's a different kind of discussion. But when it comes to that case, I think we need new rules. For example, in a family law case, when you divide the assets between the husband and wife, there are a couple of models that they use to put all the numbers into the computer and they dispute cell, certain arrangements. Of course, the judge can accept it or not, but it's immensely helpful in determining what the outcome would be. I guess the AI going forward would be a more advanced form of aiding the determination. I think the accountability part should be worked out. But again, when it comes to the institutional arrangement and the revolutions, then we definitely need new legislations. I think that those are things that are kind of cropping up in the United States and also being studied in Korea as well. When it comes to maybe the transparent usage of artificial intelligence, AI can make legal research and duck and drafting far more efficient. So where should courts, if at all, draw the line between responsible assistants and excessive reliance? Well, I think so far that the court can say that, well, AI, you don't have to band AI all together. AI can be a very useful technological tool, but the line should be drawn that whoever signs that paper, whoever is responsible as a legal professional or advocate or lawyer should should be responsible. I think that the line is pre-clear. And if you look at all the cases, the lawyers get sanctioned, the lawyers get fined or disbard. The institutional arrangement is pretty clear. The lawyer is a centerpiece of this profession, especially in the United States, where all the bar associations are a private bar. And that makes sense, because there are safety guards. If the lawyer makes mistakes, there are malpractice lawsuits that they can-- the clients can use or the sanctions that the court can use or the state bar of policies themselves in it. So I think the infrastructure is already in place, even with a fast development of AI. But again, as I said, if the AI replaces lawyers or places judges to any certain significant respect, I think the new rules and new paradigm should come into place. But what we are seeing so far is that the AI is a very capable technology that happened or appeared numerous times before. The computer propping up telephones and other technologies that are very, very helpful and threatens the profession. But the legal professionals have survived. And I think for the time being, that will continue to be the case. Because we still need human touch, right? Strategy, ethics, negotiation, empathy. I don't think this is something that artificial intelligences even close to mimicking. And I think that's a key word, too. It would want to ideally mimic the best human lawyers and still has a lot of room to grow. It seems that you've already highlighted the same point multiple times again. The key point is that no matter what the invention of technology, lawyer signature on a filing represents that the lawyer has made a reasonable inquiry and stands behind its accuracy. But I guess I'll just ask one more time on the possibility of ground rules. Should there be still more ground rules, should courts establish rules or guidelines on AI use in legal proceedings rather than leaving the issue up to individual lawyers and the law firm and the judges? To say more about that, even though the rules in place, I think there is a need to strengthen it. Because it's so easy to use the AI or chat GPT or a clause and come up with a very assuring sounding or logically phrased the court documents. I think that's why even if you read the outputs, unfortunately you've got to examine it closely because you might just say, hey, this sounds so good. And the bypass all the wrong citations and all those. So I think there's a genuine need to strengthen the role. But I think we can go more than that. I think that this time around the AI becoming a kind of institutional tool is something that will fundamentally change the US structure. I think that both sides and both sides. I think the AI can make the accessibility of the law and the legal services more readily. I mean, still, the attorneys fees are pretty high for many individuals. So that can make things easier for the everyday, the societal matters be connected to the legal analysis. But at the same time, you have to have the human touch to it. I think one level I would add aside from the attorneys is that historically we had the legal development, which is in the legal side to it in England. And then there is a notion called equity, which means that even if the law says certain things, if you look at it from the human perspective, human's point of view, then there are things that people need to get more sympathy or get more departments, get more services from the authorities, even aside from what the book says. So the notion of law and equity kind of beginning together to achieve the fairness and the human form of the legal justice has existed so far. And I think it shouldn't change. And when it concerns, I guess, this larger question of future of law, looking at Professor Song, how do you expect AI to perhaps reshape the legal sector? You mentioned the good and the bad, of course. And which human skills then will become even more important? You've already highlighted empathy, ethics, it seems. So what-- Airplaceable human skills, we come most important in the future of law. Right. I think there are a lot of speculations about where do we go or where do we end up. I think the conservative approach would be, if you look at the history, we had the printing press typewriter, computerized database, the electric discovery. Those things have arrived and helped, actually, the legal professions. So those are the things, again, the analogy would be very capable legal assistant, or a very capable tool that would help us. I think the more revolutionary or disruptive future would be if it replaces the legal profession or function in a very significant way. As I mentioned, the legal tools that we use in the family of law could be expended. And also, we might have, just like in the baseball, umpires, they use the technology to just strike the ball these days. I don't know if all the baseball fans really like that. Right. So in that sense, it could be more precise, but at the same time, the law-- the bond is born out of social arrangement. And our society and our life and what happens is not exactly always fair or logical. So putting the exactitude of the AI towards a somewhat faulty society that we have to manage with the human beings cooperating together, that would be a very big and pretty big project to put in the AI system over what we have as a judicial and legal system. I'm pretty pessimistic that the lawyers will go away, or the judges will go away, or the justice system will be entirely taken over by AI. Maybe it's coming from a bias view of the legal profession. I think not only the hopeful statement, but I think what it should be is that the human beings and our responsibility and our conscience and our sense of empathy and the sympathy should be the basis of what we call justice. And can artificial intelligence actually contextualize all those aforementioned complexities of human societies, human emotions, and human experiences for short answer is no. And I think that there's a consensus around that. I don't think that's different by your biases. Thank you so much, Professor Song, for joining us. We'll speak to you again soon. My pleasure. Thank you. If you're listening to our program using the podcast service, just a reminder that we do go live Monday through Friday 7 a.m. career standard time. So tune in and help us make the show more informative by giving us your input. See you bright and early on Good Morning Soul.