Episode 22: May 5, 2026 | Cornered: Jonathan Nessler on AI and the Future of Lawyering

Transcript

00:00:01 you're an attorney?

00:00:02 I have a friend who...

00:00:03 I've been meaning to update my will, but I just bought a new house.

00:00:06 I want to start a business.

00:00:08 My brother was fired.

00:00:09 Hey, my friend got divorced a while back.

00:00:12 I just don't understand how her ex...

00:00:14 You've been there.

00:00:15 At a social function, meeting friends of friends.

00:00:18 Word gets out that you're an attorney and suddenly your night is filled with partygoers asking you quote unquote simple legal questions.

00:00:25 The questions are seldom in your area.

00:00:27 Some of the stuff you haven't thought about since law school.

00:00:30 You're being cornered out of court.

00:00:33 In the Cornered Out of Court podcast from IICLE, you'll hear from fellow attorneys about the questions they get and the responses they give to escape being cornered.

00:00:46 In a short amount of time, generative AI has moved from the margins of the average person's awareness to all of us witnessing industry leaders and consumers shift dollars and human capital because of the technology's impacts.

00:01:02 Jonathan Nestler has published a new book about how the practice of law is forging novel paths as a result of widespread access to generative AI.

00:01:13 My name is Jonathan Nessler.

00:01:14 I am a trial lawyer in Springfield, Illinois.

00:01:17 I work with the Law Offices of Frederick W.

00:01:19 Nessler and Associates Ltd.

00:01:21 It's a family firm.

00:01:22 I've been there practicing for, hard to believe, but almost 20 years now.

00:01:28 I do mostly plaintiff's trial work.

00:01:31 The book is titled Infinite Counsel: A Lawyer's Guide to Understanding Artificial Intelligence and the New Economics of Practice.

00:01:40 What prompted you to write it?

00:01:42 I've always had a deep interest in technology ever since I was young.

00:01:48 I remember my dad used to bring his old office computers home and one day he brought home a Mac Classic that was the first computer I ever sat down and used and was just immediately enamored by that.

00:02:02 And so I've always had a deep personal interest in technology and as I went through my life that continued into practice.

00:02:10 One of the things I focused on when I first started practicing was using technology back 20 years ago, much different tools, but using technology to augment the practice of law.

00:02:21 And there's always been, I think, an apprehension in the law to use new technologies because we have so many restrictions.

00:02:30 And we're seeing a lot of that with AI, especially confidentiality, attorney client privilege, things like that, don't have a natural fit with.

00:02:40 For instance, cloud storage, just as an example, you're taking confidential information, you're storing it on somebody else's server.

00:02:46 How do we reconcile that with sharing communications with a third party?

00:02:51 So it wasn't always an easy task to find tools that were okay to use under those types of restrictions, but it was always kind of fascinating to me.

00:03:02 So fast forward to the release of ChatGPT when OpenAI first made their GPT models available to the public, I started using those and immediately realized that this was going to change the practice of law.

00:03:19 I mean, the fact of the matter is that in those days, it was just text generation.

00:03:25 We hadn't quite gotten to reasoning yet, but the idea that you could generate text at scale was going to immediately impact the practice.

00:03:34 And so I started using the tools, talking about them, writing articles about them almost immediately.

00:03:40 And what we've seen since then, in the past three or four years, is just exponential growth in their capabilities.

00:03:48 The first models could generate text, and it was pretty good, but there was a lot of problems with it.

00:03:54 And then we moved into reasoning models, and it became clear that not only were these models capable of producing text that looked like a lawyer wrote it,

00:04:02 but were also capable of thinking like a lawyer, reasoning like a lawyer.

00:04:07 And although that may seem pretty astonishing, I would ask you to take into consideration for a moment, by the time we move from GPT 3 to GPT 4, GPT 4 passes the standard bar in the top, I want to say 95th percentile or something like that.

00:04:25 In other words, it's beating 95 percent of bar applicants at the grade that it's able to achieve on the bar.

00:04:34 And that is GPT 4.

00:04:36 That's two years ago already.

00:04:38 So it again became apparent that soon we were going to be living in a world where we have machines that not only can write like lawyers, but can think like lawyers.

00:04:48 And in my mind, we've already entered that world.

00:04:52 And so as I was thinking about those things, writing about those things, and giving presentations about those things, I thought, well, I might as well write a book.

00:05:00 You know, we talked before the podcast about the fact I was a communications major and undergrad with a journalism focus.

00:05:08 And so I've always really liked writing.

00:05:12 I've always wanted to write a book.

00:05:13 Frankly, I wanted to write a fiction book most of my life, but I guess nonfiction was the path that I ended up taking.

00:05:18 So I put a lot of those ideas in this book, Infinite Counsel.

00:05:24 What does the phrase infinite counsel describe?

00:05:28 Infinite counsel is the concept that we're quickly moving into a world where the supply of competent legal assistance is going to approach infinity.

00:05:39 And what I mean by that is once we have machines that are capable of providing competent legal assistance,

00:05:47 You have to think about it in kind of the way we're using the tools now.

00:05:49 It's not like there's going to be one AI as such that is providing legal assistance to millions of people.

00:05:56 It's going to be an instance of that.

00:05:58 So anybody can spin up an instance of a very competent legal mind and be able to ask it questions and be able to seek counsel from it.

00:06:08 And once we approach a world, and we're probably pretty close to it now,

00:06:13 where there's an infinite supply of legal assistance, what does that mean to the practice of law?

00:06:19 What does it mean when anybody can go to their computer and access an AI that's capable of providing, in most cases, better legal assistance than they could get from a lawyer?

00:06:32 I think there's a lot of really good benefits to that.

00:06:34 You know, first of all, there's this justice gap that's going to be met pretty quickly because

00:06:40 as we've seen using the tools now, they're really cheap to use.

00:06:43 But there's also a lot of downfalls.

00:06:45 They're cheap because they're taking the information that people are putting into it, they're using it to train, they're using it to sell products.

00:06:52 And obviously there's that same conflict that we talked about with the cloud storage that if people are asking for legal advice from a machine without understanding that

00:07:05 there's somebody on the other side of that, a company, whether it's OpenAI, whether it's Google, that are taking the information.

00:07:12 It's not confidential, it's not privileged.

00:07:15 And really, we just saw a case come down, I think it was last week, that talked about that concept.

00:07:26 A judge just ruled on that.

00:07:27 I believe that was out of New York, Southern District.

00:07:30 And the opinion was clear that those are subpoenaable.

00:07:34 They can be used against the person because they're a communication with a third party that's not a lawyer.

00:07:39 There's no privilege there.

00:07:40 So we have a lot of problems that are associated with it, but the book really focuses less on those problems and more on what that means for law firms.

00:07:50 You know, what does it mean

00:07:51 when we remove the intelligence barrier that's always existed between lawyers and the rest of the world.

00:07:58 Lawyers have always, as part of their licensure, held a special knowledge, a special ability that they learned through their education, through their experience practicing law, that was really the product that they sold.

00:08:12 It was the ability to understand the law, read the law, apply the law.

00:08:18 And because of that intelligence barrier, we were able to justify, for instance, billing hundreds of dollars by the hour.

00:08:25 We were able to justify selling a service that was otherwise a very scarce resource.

00:08:32 And what we're seeing through models that are capable of providing competent legal assistance is the removal of the intelligence barrier that set lawyers apart.

00:08:44 And so that's really the concept that the book explores.

00:08:49 What do you think AI tools are effective at helping lawyers with today?

00:08:53 And what do you expect these tools will be able to assist with in the future?

00:08:59 Yeah, so I think that's a really important question.

00:09:02 What the tools are useful for today

00:09:05 involve essentially paralegal tasks and first year associate tasks.

00:09:09 If you use the high end models, and I think it's an important point to make, that if you're using the free model through OpenAI, through Google, through Grok, that's the XAI's model, you're not using the most powerful model.

00:09:25 If you have the subscription to any one of those platforms, you're using a much different model.

00:09:31 So

00:09:33 I think at the outset, we should distinguish between the free models and the pay for models.

00:09:37 If you're using a pay for model, that's the cutting edge or frontier model, as they're called, through any one of those platforms, those models are capable of performing legal tasks at or probably now better than a paralegal or a first year, second year associate.

00:10:12 The types of tasks that you would give to a paralegal or a first year associate.

00:10:18 Things like research this legal issue for me, or take this case and distinguish it from the facts and the matter that we're handling, or write a brief, write a response or a reply.

00:10:32 In the current moment, the frontier models are highly capable at those things.

00:10:40 More than just highly capable, they can probably do it better than most lawyers, especially when you talk about the fact that you could get the draft done in 20 minutes as opposed to two weeks.

00:10:57 And if you don't like it, it's 20 minutes away from another draft, 20 minutes away from another draft.

00:11:01 You can ask follow up questions, you can explore why it decided to go certain routes, and it can explain itself instantaneously.

00:11:09 You know, there's no delay like you would get from a human associate.

00:11:15 And so I think that we've already started to enter the world where we will start to see paralegals and first year, second year associates replaced by artificial intelligence.

00:11:27 But we're still dealing with, in today's world, narrow artificial intelligence.

00:11:31 And so it's an important distinction to make to answer your second question.

00:11:35 Narrow artificial intelligence is an artificial intelligence designed to do a specific thing.

00:11:42 And so we see clear examples of that with maybe 20 years ago, 15 years ago with an artificial intelligence designed to play chess.

00:11:53 And those artificial intelligence started off being able to beat some players, went on to be able to beat most players, and now can beat all players.

00:12:04 So you have this artificial intelligence designed to play chess that can beat any player at chess, but couldn't beat somebody at checkers.

00:12:11 You have an artificial intelligence designed to play Go.

00:12:17 It was kind of that same era.

00:12:25 It's an Asian game that was very popular, also considered to be more difficult than chess.

00:12:31 And you saw that same timeline.

00:12:35 Better than most, better than almost all, better than everybody.

00:12:38 But still couldn't beat somebody in checkers either.

00:12:40 Because they're narrowly designed.

00:12:44 They're designed to do one specific thing.

00:12:47 That's as opposed to artificial general intelligence.

00:12:51 Artificial general intelligence is an intelligence that would theoretically, because it doesn't exist today, be able to think and act like a reasonably capable human under all circumstances.

00:13:02 I think the only way to answer your question is in a world where artificial general intelligence exists.

00:13:09 Once artificial general intelligence exists, it will be capable of practicing law at or better than a reasonably capable attorney at first.

00:13:20 And then we saw in the examples of the narrow artificial intelligence, it's going to accelerate to a level where it's going to be able to practice law better than any lawyer on the planet.

00:13:35 And so that is a totally different world.

00:13:39 It's a world where an artificial intelligence, a machine, is capable of out thinking any human lawyer at any task at scale.

00:13:50 So theoretically, you could take the information gathered from a client at an intake, provide it to the artificial intelligence, and without any further prompting, without any further tasking, it would

00:14:02 provide you with whatever the next step was, what the course of the case is going to be, case plan, et cetera, and all the materials related to that.

00:14:09 It would literally do the entire case with human supervision for now.

00:14:15 And that is going to fundamentally change how lawyers practice law.

00:14:20 And we're already seeing it through the narrow artificial intelligence where we're starting to see, as I mentioned, paralegals and young associates being replaced.

00:14:32 Because it's frankly just more capable, cheaper, and it's the option that on the business side of the practice of law people are going to opt to take.

00:14:39 Hiring an associate for seventy five thousand dollars a year versus using a technology that provides you with ten associates for five hundred dollars a year or one thousand dollars a year, it's kind of a no brainer.

00:14:53 It's one of those things that from a strictly economic standpoint law firms as businesses aren't going to have any choice.

00:15:00 In the introduction, you describe how lawyers as problem solvers are well positioned to tackle challenges AI presents to knowledge workers.

00:15:17 How do you see lawyers demonstrating a path forward for other professionals?

00:15:23 To answer that question, we have to go back to kind of the primary topic of the introduction and really the book as a whole.

00:15:30 And the introduction is titled, Is This Time Different?

00:15:34 What I mean by that is historically, when humans have created tools, they've created something that was meant to mitigate a human weakness by leveraging the primary human strength, which is our intellect.

00:15:50 The reason that humans are the apex predator on the planet is because they're smarter than all of the others.

00:15:55 They're not faster, they're not stronger.

00:15:58 And so when we look through, for instance, the Industrial Revolution, what we see is the production of a series of tools that are designed to leverage human intellect and thereby mitigate human weaknesses.

00:16:13 Whether it's a tool designed to dig ditches faster than a human, cut down trees faster than a human.

00:16:19 Always the focus of the tool was the fact that humans had the intellect to leverage, to use the tool to do the thing.

00:16:26 And that really carried all the way through the digital revolution.

00:16:29 A computer, historically speaking, wasn't capable of thinking on its own, but it was a tool that humans were able to use their intellect to leverage the capability of.

00:16:38 An example that I used when we were talking before the podcast was if you gave a monkey a computer, they would probably bash it into a tree.

00:16:49 They wouldn't use it for the design of the tool, and there's nothing the computer could do.

00:16:53 Like I said, it's just a tool.

00:16:54 It doesn't think on its own.

00:16:56 Humans use their intellect to leverage the tool.

00:17:00 For the first time, we're designing a tool that can replace the one thing that sets us apart.

00:17:07 And I think that is the point that lawyers need to give a lot of thought about.

00:17:13 Obviously, if you have a law degree, you're not just practicing trial law in Springfield, Illinois.

00:17:18 You're also going into creating laws, you're going into writing laws, you're going into interpreting laws in the case of judges.

00:17:25 And one of the ways that lawyers can play a role in this is start to try to develop frameworks of what society is going to look like once humans have been replaced as the primary source of intellect.

00:17:38 Once there's no longer a barrier to the amount of intelligence that exists in the world.

00:17:47 Because that's where we're moving to.

00:17:48 We're moving to a world where intelligence is no longer a scarcity, but instead an abundance.

00:17:56 And the changes that that's going to bring are so dramatic that it's really hard to imagine.

00:18:03 And frankly, I don't know that I have the intelligence to imagine a world where we have machines that are not just capable of thinking like a reasonably capable person, but thinking orders of magnitude higher.

00:18:16 And that's really the world where we're headed.

00:18:20 Because once you have an artificial intelligence that can think like a human,

00:18:25 you can set that artificial intelligence towards creating a smarter artificial intelligence.

00:18:30 And once that's created, then a smarter and a smarter and a smarter.

00:18:34 And soon you have machines that are not just smarter than any human, but are capable of thinking at a level that humans can't understand at all.

00:18:43 I just, I don't see how I can reconcile that in my mind with humans continuing to have a place in employment.

00:18:53 And I know that sounds extreme, I do, but at the same time, it's not a certainty.

00:18:59 But it might be different this time because what we're creating this time is fundamentally different than what we've created before.

00:19:08 We're creating something that replaces the one advantage that humans have always had, our intellect.

00:19:14 And once that's replaced, there's no going back.

00:19:20 I guess the line that I draw in the book or the example that I use is education.

00:19:24 Something I know you're passionate about.

00:20:00 And that is, education in the United States is free from kindergarten through twelve.

00:20:10 We pay taxes, but fundamentally it's free.

00:20:14 We can send our kids to a public school.

00:20:17 And although that resource is free, there's still a number of people, myself included, that opt to send their kids to a private school.

00:20:28 And what private schools have done to exist in a world where the commodity that they provide, education,

00:20:35 is otherwise free is they've come up with ways to set themselves apart.

00:20:39 They've come up with value adds that people are willing to pay for.

00:20:44 Those value adds could be a religious component.

00:20:47 Obviously send your kids to a Christian school or a Catholic school.

00:20:51 A smaller classroom component.

00:20:56 Selling themselves as providing a better education.

00:21:00 A more safe environment.

00:21:03 And all of the other things that you think of when you think of why is it that people are paying to send their kids to a private school.

00:21:14 When they could otherwise send them to a public school.

00:21:20 I think that as lawyers, that's one of the approaches we're going to have to take.

00:21:24 What does a human lawyer offer that the otherwise free variant of intelligence doesn't offer?

00:21:32 Appearing in court comes to mind.

00:21:35 That is a really low hanging fruit.

00:21:38 Obviously, these artificial intelligences aren't embodied.

00:21:40 Licensure is another one.

00:21:43 The capability of being able to review what's been produced to make sure it's accurate.

00:21:47 And make sure that someone's held accountable if it's not accurate.

00:21:50 And we're seeing that throughout the legal profession now.

00:21:54 With lawyers who have filed briefs that have hallucinations.

00:21:57 You know, it's not the AI that's being penalized for that.

00:21:59 It's the human lawyer.

00:22:02 Because the human lawyer failed their supervisory or oversight role.

00:22:07 In making sure that what they were filing with the court was not only accurate but was true.

00:22:14 And in some of those cases, the cases that were cited were hallucinated.

00:22:20 Which is a fancy way of saying the AI made them up.

00:22:23 And so in other words, it's not that they filed a brief that they used AI and the AI made it up.

00:22:28 They filed something that was inaccurate and false with a court.

00:22:33 And they signed their name to it.

00:22:38 And so obviously those two things are value adds.

00:22:40 But again, as we move into a world where artificial intelligence is capable of out thinking us as humans.

00:22:49 As a profession, I think we need to take a step back.

00:22:54 And really consider what that means.

00:22:58 Is it something that we want as a legal profession.

00:23:03 To have machines doing ninety nine percent of the legal work.

00:23:07 While humans are merely providing oversight.

00:23:11 One example that I heard very early on was that soon

00:23:17 We're going to move into a world where someone generates an email that's two pages long using AI.

00:23:33 Send it to their coworker who's going to use AI to minimize that email into five points.

00:23:42 Read those five points.

00:23:45 Then generate the response email back.

00:23:50 And it's going to be two pages long.

00:23:53 But the recipient's just going to again minimize it to five points.

00:24:00 So you have all of this text that's being written by these machines back and forth.

00:24:07 And no one's reading it.

00:24:10 They're just using machines to interpret it.

00:24:14 And I thought to myself, the parallel to that in the law.

00:24:21 Is a lawyer generating a brief.

00:24:27 The responding lawyer taking that brief, putting it into a machine.

00:24:33 Generating a response.

00:24:37 And then generating a reply.

00:24:42 And then filing those with the court.

00:24:55 The judge taking the brief, the response, and the reply.

00:25:00 And then generating a reply.

00:25:04 And then filing those with the court.

00:25:07 The judge taking the brief, the response, and the reply.

00:25:11 Putting it into a machine and generating an order or a judgment.

00:25:20 All the while no human being ever really reading through the material.

00:25:27 But instead just relying on the machines to do all of the work.

00:25:33 And obviously that's not a world we want.

00:25:35 So as attorneys, I think that when I say we're uniquely suited to address those issues.

00:25:43 First we need to recognize them as possible.

00:25:47 I think there's an extraordinary amount of denial right now.

00:25:54 That they're never going to be able to do this.

00:25:57 I clearly remember three years ago people saying they can't reason.

00:26:02 They're never going to be able to reason.

00:26:04 And then two months later, reasoning models come out.

00:26:10 And two years later, reasoning models are so extraordinary.

00:26:16 That if you're using them right, they've almost eliminated the hallucination problem altogether.

00:26:24 And a lot of these entrepreneurs on the frontier say.

00:26:33 You're currently using the worst version of ChatGPT that you'll ever use.

00:26:44 And I think that's a really interesting way of saying it's only going to get better from here.

00:26:49 And we can keep denying it.

00:26:52 Or as a profession, we can look at what's coming.

00:26:54 We can explore the possibilities.

00:26:58 Explore the limitations that we should put on allowing machines to essentially practice law.

00:27:03 And take maybe even a defensive posture to protect the profession.

00:27:07 Because at the end of the day, being a lawyer is a really hard thing.

00:27:11 And I don't think many people are going to be willing to do that job for twenty five dollars an hour.

00:27:17 It's just not something that's going to be a feasible employment opportunity.

00:27:23 And if these machines are providing legal assistance at scale.

00:27:29 What other world exists?

00:27:33 Because at the end of the day, if somebody is able.

00:27:39 To obtain competent legal assistance for pennies an hour.

00:27:43 Why would they pay hundreds of dollars an hour for that product.

00:27:50 In the absence of either value adds, like we discussed.

00:27:54 Or taking some defensive posture to protect the profession.

00:27:57 To make sure that there's human involvement.

00:28:03 And I think that human involvement in the law is an extraordinarily important thing to protect.

00:28:10 Thank you, Jonathan.

00:28:12 Jonathan Nessler's book, Infinite Counsel, A Lawyer's Guide to Understanding Artificial Intelligence and the New Economics of Practice.

00:28:23 Is available from integratedcognition.com and at other online retailers.

00:28:31 If you have an idea for a topic you would like to hear discussed on the IICLE podcast, we welcome your suggestions by email.

00:28:42 Our address is info at iicle.com.

00:28:47 IICLE is a 501 c three not for profit based in Springfield, Illinois.

00:28:55 We produce a wide range of practice guidance for Illinois attorneys and other legal professionals in all areas of the law.

00:29:11 With the generous contributions of time and expertise from attorneys, judges, and other legal professionals.

00:29:24 If you are interested in our many authorship and speaking opportunities, please give us a call.

00:29:31 At two one seven seven eight seven two zero eight zero.

00:29:36 Or visit the Contributor Resource Center at iicle.com slash contributors.

00:29:45 Thank you for joining us for another edition of Cornered Out of Court.

00:29:50 Brought to you by the Illinois Institute for Continuing Legal Education.