Are the bots coming for our jobs? Will law firms crate when AI erodes the leverage model? Law professor Jonah Perlin advises us to keep calm and bill hours: AI is but another technology to add to the legal toolbox. In this episode, Perlin explains that time is but one element of the billing ecosystem. In this episode, we talk about the value of legal services and how lawyers and law firms can demonstrate value (and adjust fee models) in the AI age. Perlin also hosts the incredibly popular How I Lawyer podcast and supports law students and early career lawyers to find career fulfillment.

—

Listen to the podcast here

    

Will AI Kill The Billable Hour? With Jonah Perlin Professor, Georgetown Law

My name is Jonah Perlin. I am a law professor at Georgetown Law and the host of the How I Lawyer Podcast. A fun fact about me is that I remember the day I got the internet in my house, and it was one of the most exciting days of my life.

—

Introduction To The Podcast And Jonah Perlin’s Background

On this episode of the show, I am excited to welcome Jonah Perlin, who is a law professor at Georgetown and host of the How I Lawyer Podcast. Jonah, you do not know this, but I followed your podcast and tried to learn from you as I was getting this show started almost three years ago. It is a real thrill to talk to you live.

That’s great, Stacy. I really appreciate it. It has been fun watching you build your podcast and the general legal podcast ecosystem grow. I also learned by listening to folks, too. That is what is very cool. I think about putting content into the world and hopefully making the world a more transparent place by putting that content in the world.

Demystify the legal department. We are going to talk about something that I think can help, and I want to demystify. We are going to get into why you are interested in it. You have written an article and done some LinkedIn posts about how AI affects law firm billing. I have my popcorn out, trying to see where this is going to go. Your article, How the Billable Hour Can Survive Generative AI, got my attention. Not that I read a ton of law review articles, but this is an evolving space. You are a law professor. You were a law clerk at one point. Why were you interested in law firm billing?

It is a very fair question. There are not a whole lot of people in law schools who are particularly interested in law firm billing. From my perspective, I got into this business to educate the next generation of lawyers on how to practice as lawyers. That is my North Star. That is what I care deeply about. I started my career as a litigator at Williams & Connolly here in DC and was a judicial law clerk as well, and had the opportunity to adjunct teach a class at Georgetown. I went to Georgetown as well.

That is a whole other story that I have told on my podcast, but I had this opportunity to adjunct teach a class. What I learned pretty quickly was that I liked being a litigator, and I think I was sort of okay at it, but I really liked teaching, and I felt like it fit my skillset even better than my day job. Immediately, that made my antenna go up and say, “Maybe I should try to find a way to do this.” I also really wanted to do it at the intersection of practice as opposed to writing and teaching about law. I wanted to write and teach about being a lawyer.

Which is, by the way, underdone at law school, right?

It is. The way, and then it has changed a lot over time. People sometimes think it is still the way it was when they were in law school. It really depends on where you go to law school. We could have a whole show just on that particular question. Thankfully, I was able to end up back at Georgetown, a place where it is a prized area of teaching, and we make it work. It got to the point when you are a professor, when you are thinking about, “What am I going to write about? What is my scholarly impact going to be?”

My interest in the intersection of law and technology goes back to my practice as a lawyer, 100%. I am of the generation I remember we just had a big snowstorm here in DC a couple of weeks ago. I was thinking back to the time when we had a big snowstorm, the Blizzard of 1996, when I was in elementary school. I remember that was the day we got the internet at our house. I am of that generation that I remember pre-internet days, but I am as close to a digital native as possible. I feel like I won the lottery by being that age at that time when the world was changing.

Technology has been on the back of my mind through everything I have done, personal, professional, etc. We do not think enough every day about how technology has changed the law, but also how the law has changed to respond to technology. We have a really rich history as a legal profession of responding to whatever is changing in society, and technology is just a piece of that.

We don’t think enough, day to day, about how technology has changed the law—and how the law has also evolved in response to technology. Share on X

That is how I got interested in this topic. Billing in particular, that is really just a product of the fact that I was reading a lot of his must be the death of the billable hour headlines, and I was skeptical. The great part about my day job is that when I am skeptical about something, I can test that hypothesis and do a deep dive. That is how that particular project sort of got started and went quickly.

It will be interesting to see what impact it has. A lot of law firm lawyers will relax their shoulders with your hypothesis that the billable hour is enduring. I wanted to probe a little bit. I am a consumer of legal services. I am a general counsel. We use outside counsel. I am not surprised, but it is interesting to me that I have not seen an impact of AI on my legal bills.

In fact, they have stayed the same or increased. When I ask firms about how they are using AI, “Let me know what I want to know, what you are using, what tasks you are using it for.” A lot of folks say, “It’s not that good. I do not use it,” and maybe it is generational, but are you hearing that from others? Maybe you do not interact with people like me all that much.

The AI Efficiency Hypothesis And Rate Inflation

No, I try to interact with people like you all as much as I possibly can. I think I would not be doing my job or studying what I study well if I did not. It is why I love having conversations like this. They become public, and people can sort of sit in on them if they want to. The immediate response I have to the question would be, what does it mean to see the impact of AI on your bill as a consumer?

There is an assumption baked into that question that the way I would see it is for my bill to go down. That makes sense. There is what I call in the article, the AI efficiency hypothesis, which is just a fancy way of saying, if you can do things faster, then one of the variables that contributes significantly to hourly billing is the number of hours that a particular task takes. If that goes down, then the compensation to the law firm should go down.

My whole article is predicated on the idea that A, maybe the AI efficiency hypothesis is not 100% right. Even if it is right, the number of hours billed is only one variable among many in what I call the billable hour equation. For example, if their rates go up to counteract their hours going down, the bill is the same to you. It is because there is a causal relationship between using AI and their rates going up.

Let us explore the cause. To me, the cause is that when you make law firm money, you want to keep earning law firm money. The rate goes up not because I am a better lawyer or because of inflation, but because the rates that I see are inflated by seven to 25% per year. It is chasing, I think, and maybe I am jaundiced here, but it is chasing the law firm lawyer’s expectation of compensation.

The billable hour, and we can talk a little bit about the history of it as well, is by definition a way of pricing legal services. The way of pricing legal services is the way that lawyers make money. You are absolutely right that the rate does not have to necessarily go up unless the lawyer wants to make the same amount of money. Let me make the case why the rate ought to go up, which is a slightly different case.

Let us hear that.

Two things come to mind immediately. The first is that you, the consumer of legal services, were paying for a lot of time and energy on tasks that were basically very low-level tasks for your lawyer to do. You are paying not just for the lawyer to come up with the great theory of the case, but you are also paying for the document to look pretty. Now, AI makes the document look pretty, and the lawyer’s time spent is on those higher-order, more important tasks. Thinking just from a purely economic perspective, if I am getting rid of the admin and the basic tasks that you do not really want to pay for to begin with, one would then assume that it is reasonable that you should charge more money because you are doing more high-impact work.

Except that the conclusion means that the hourly rate is sort of a blend. Like, I charge $1,000 an hour because some of my time is spent on editing documents, and the other time is coming up with a strategy for a case. I do not know that that is true.

That is the fundamental question. The fundamental question is, what are these rates tied to if they are not tied to an articulation of what you, the consumer, are willing to pay the lawyer to do? Think about it this way. Think about it as the idea that different people make different hourly rates in regular hourly work. We charge a different hourly rate depending on how hard or how much education the activity requires. You are absolutely right on it being an implicitly blended rate.

The second reason that I think lawyers can justify arguing to their clients that their billable hours should go up to compensate for the number of hours going down is to let us be clear about what consumers of legal services are buying. They are not buying hours. They are buying outcomes. Hours are a way for us to get back into how we price those outcomes. Ultimately, my colleague, Mitt Regan at Georgetown, wrote a book called BigLaw, where he sort of studied big law firms empirically.

He and Lisa Rohrer wrote this book, and it came out in 2019, right before the pandemic, unfortunately, because the world changed a little bit. What is really interesting about their empirical studies and their qualitative discussions with practicing in-house counsel are the counsel often said, “I actually do not care what their billable hour number is or even how many hours they spend. What I care about is the cost of what I value this task to be. If they are going to charge me more than I value this task to be, I am not going to hire them, and if they charge me less, I feel like I am getting a deal.”

The hours then become frankly just a way of having that negotiation in advance that protects both the client and the lawyer. Plus, we know that clients, when lawyers charge flat fees, are still asking lawyers to see their hours so they can sort of justify that work. By contrast, when lawyers are putting out bids, they are essentially working backwards from what they think the client is going to cost. Essentially, the billable hour really is just a proxy for the negotiation between in-house counsel and outside counsel.

It is a proxy. It is ultimately a driver of the cost, because whereas you do a flat fee, we are both taking risks.

It is de-risking.

It is de-risking except that the consumer, like, your estimate of the hours “should take” is based on your experience having been a law firm lawyer, and mine is dated. I have been in-house for 15 to 20 years. As I think back, there are resources, there are processes, technology, etc., that were not in place when I was practicing. I have a skewed vision of what the “should” is.

That is what I mean by it is a proxy. We are using billable hours to sort of back into a negotiation about what the cost of a particular matter for a particular client at a particular time is. We have seen this in other industries throughout history that have sort of become more automated, with less manual time and effort. What you often see is that the quality that is desired by the client goes up. Let us use spell check, for example.

If you got a brief today that was not spell checked and had a bunch of typos, you would go crazy the second that happens. In the age of typewriters, when that just happened all the time because of the technology available, you may not have thought about it in the same way. That is a very basic example, but the quality goes up, and the expectations go up. I remember hearing from a law firm partner when I was a very junior lawyer, talking about what it was like to be a junior lawyer when he was a law firm associate.

It was, “We’d write our brief draft by hand. We’d give it to the typing pool and go play golf. The next day, we’d come in, we’d edit it. We’d give it back to the typing pool. We’d play golf. We’d send it to the client. Five days later, we’d get one set of hand edits, and then we would file.” That is not how work is done today. The quality and the amount of work and time that is spent on any given brief, even if it is the same brief that it was 50 years ago, have fundamentally changed because the technology behind it has changed.

That is interesting. That gives me a little bit of comfort that, yes, while AI is a disruptive technology, we have been there before.

That is my whole point. There are a lot of people who disagree with me on this who believe that AI, AI exceptionalists who believe that AI is so different. I recognize that AI works differently and does different things. At this point, I still believe it is the next chapter in a story where the legal profession is changing. I actually think people like to focus on, for example, hallucinations. I actually think, on balance, the quality of lawyering across the profession will likely go up as a product of the use of AI, not down.

AI works differently and does different things. At this point, it is the next chapter in a story of a legal profession that is changing. Share on X

Hallucinations, Factual Inaccuracy, And Lawyer Credibility

Since you mentioned the H-word, hallucinations, I do want to talk about this. We are going to get to it later in the conversation, but I cannot help myself. You see stories about this. There are people getting sanctions, even judges are using it and getting burned, and people are not site-checking their filings. When we talked in prep, I was sort of aghast, sharing my real concern about that and worry about the lawyers that work for me. You pointed out that this has always been a problem based on your experience as a law clerk. Help us understand why, yes, hallucinations are a thing now, but it was maybe always happening?

Ultimately, let us just be clear that, as a legal research and writing professor, hallucinations also horrify me. It almost feels like I have not done my job if a student goes out into the real world and submits a brief that has incorrect information on it, or any document, but I am just using briefs as an example. There is someone on the internet who has been tracking hallucinations, I think hallucination cases, so sanctions cases basically. We are up to over a thousand. We have crossed that threshold. We crossed the thousand hallucination mark.

A couple of things. First, we only see the errors. We know, for example, that according to some statistics, more than 50% of lawyers today across all age demographics use AI-assisted tools at least once a week for their profession. Again, that statistic may be wrong, but just one statistic that I have seen. If we’re only seeing a thousand hallucinations tells me that there are hundreds of thousands of briefs or millions of briefs that are getting filed using AI that do not have hallucinations, or we are not catching them. Those are the two options. Either way, we do not have a good comparator.

When you only know the numerator and not the denominator, it is frankly a misleading number. It comes down to, I think, the fact that lawyers believe we can never make mistakes and therefore the denominator is irrelevant. This is where my second point that we were talking about comes in, which is this idea that we did not put factual inaccuracies in briefs before AI is just wrong. The type of inaccuracy may have been different. We did not usually make up case names, I concede that 100%.

Actually, I think the harder thing to identify, especially as a judge, is when a lawyer talks about a case and basically fudges the case either intentionally or otherwise, to make it look better for them. If the opposing counsel, and we know three-quarters of civil cases do not have lawyers on at least one side of the V, does not catch it, and the judge does not catch it, you have these situations that end up happening where law is made based on false precedent articulations.

You saw that as a clerk?

Yes.

A lot of times, like a thousand times?

Not all the time, but it is what you check. Ultimately, I got into the habit of checking all of the important citations at the very least, especially when I was at the Court of Appeals, when we were only working on a handful of cases at any given time. It was my duty, I thought as a law clerk, to go and look and see, did the case actually say that? It is not just the law. It is also a factual record. Think about the size of the factual records in many cases.

When people make generalizations, I teach my students who are writing their briefs now not to give what I call opinions, which are articulations of what the facts say as opposed to what they actually say. We are trained as lawyers to walk up to that line where that line is to provide the necessary zeal to our client, and people cross that line all the time. In some ways, hallucinations worry me less than those kinds because they are easy to catch. If you just spent any time looking up, is there a case with this name, with this citation checked? It is done. Those are far scarier to me.

Frankly, AI does that too, because it is just summarizing, and it is trying to summarize as humans summarize, and human summaries are not 100% accurate. Ultimately, my view on this is that the sanctions decisions are helping us create an infrastructure as a profession about what the expectation is. I also think that law is practiced so fast now that I saw this when I was practicing, and I have not been in practice for almost ten years at this point, but the infrastructure we set up around, for example, site checking and verifying, had started to fall a little.

The idea that I remember I was getting edits as a junior associate to a brief from partners within minutes of when it had to get filed. You cannot site-check in that environment. There is actually going to be a Renaissance of firms that used to employ full-time site checkers. They do not do much anymore, but I think they will start doing that.

As a consumer, because I think we talked about this before, I cannot imagine anything worse than having a firm on your client’s behalf, filing something with a court or a regulatory body that had false information, false representations, or a fake case. Your credibility is completely undercut. I tell all firms, if you do, I expect a human to site-check everything in any filing you have for me.

Let me just say, here is the last piece of it, which you might disagree with and might be a spicy take, but that is what podcasts are for. I actually think that the assumption that humans are better at site checking than computers, if it has not already gone away, will go away soon, because ultimately, humans make mistakes. We know that. Humans read things quickly, and they forget to highlight the one case that they were going to check.

AI-based tools, I truly believe, are going to be both part of the problem, but eventually part of the solution, where there is going to be a one-click way in the major market tools to say, “Do all these case names exist, yes or no?” Ultimately, again, it is creating this larger infrastructure. My next article, I think, is going to be on this topic, so I have been doing some early reading on the history there. There is a really interesting history of the Shepard’s product. For those who are too junior to know what that is, you might have heard the term Shepardize.

From spicy to snoozy, I have to say.

The history of legal tech is not snoozy to me, but I will tell you this. Shepard’s, for example, was a direct response to the fact that lawyers did not have up-to-date information about what remained good case law, so they were citing case law to courts that was false in the sense that it had been overruled. What did we as a profession do? We created a bunch of tools. Shepard’s is the most successful of these tools. Interestingly, the guy who created it, Mr. Shepard, was not a lawyer. He was a law book salesman from Illinois who said, “I know I can sell this new product.” He was like the first legal tech product. AI is going to be part of the solution, not just part of the problem.

The Billable Hour As A Proxy And The CHARGE Equation

I look forward to that. I do not know enough about the tech to argue with, to say whether it is a spicy take or not, but that is my analog solution to the hallucination issue. Getting back to the billable hour issue. Maybe you could go through. I have never heard anyone talk about law firm billing in this way. What are the different variables that go into the equation?

I do not think that anyone has used these exact terms, but I think anybody who has been in the billable hour space knows them to exist. There could be other terms. It is, by definition, an oversimplification. The variables or the levers that I came up with are the compensation to the law firm is equal to the hours worked times the rate at which you charge for those hours, minus any what I call granted reductions, agreements between the Adjustments, minus expenses, the expenses and overhead of the law firm.

Why is that helpful? It’s because it spells CHARGE, thank you to ChatGPT for helping me come up with the acronym. Those are sort of the major variables that I talk about in the piece and that I think lawyers intuitively think about when they are thinking about billing. My whole point is we need to talk about how AI is going to potentially affect all of those variables, not just hours. Ultimately, if we do not have that conversation, we do not have a complete picture of whether or not the billable hour works or not.

We need to talk about how AI may affect all of those variables—not just hours. Ultimately, if we don’t have that conversation, we don’t have a complete picture of whether the billable hour works. Share on X

To be clear, my article comes off like I am saying it should stay. My main point is actually that I think it is likely to stay. Whether or not it should is like a smart econ or a moral judgment. I truly do not have a dog in that fight. The idea was that I was responding to what I saw as the prevailing argument of the time in which it was being written. Maybe it has changed in the six months since it came out, which was. Hours go down, billable hours die, we are all doing alternative fee arrangements now, and they are no longer alternative. Maybe that is an oversimplification, but that did not seem exactly right to me.

I am eager to see what the impact will be. One of the things I have on my mind is, once e-discovery came to be, and you have these big databases and Relativity and all these different tools that help law firms manage and search big, big caches of data, we all now see data charges on our bills. My prediction, and maybe this is happening already, is that we will start seeing AI tool bills as well. If you want us to use Harvey, if you want us to use something else, then you are going to pay for it.

They are going to bake that into their rates. There is another reason we did not talk about why rates go up, because the tools function to make those rates go up. There are a lot of different ways.

Rates go up because people will pay them. That is, honestly, I think that law firms put out a flyer, “Can I get $3,000 an hour?” Someone pays it, and then that sets the market.

That to me tells me, as long as that is true, the billable hour will survive.

I think you’re right.

It’s because it is not rational. The billable hour has not entirely been rational. That is, I guess what I mean. I do think you’re right. I think you are going to see that. Some people are going to push back against it.

Which you should do about data charges, by the way, the law firm will adjust those off again.

Again, I hate to be a broken record on this, but this is again not the first time that law firms have tried to pass tech costs.

Faxing charges, you were charged for postage.

When I started at the law firm, we still passed. I do not know if I am telling secrets, but I think I could say this. When I started at the law firm, many clients still paid for our legal research bills for their matters. By the time I left, many of them had negotiated that that was a cost of the firm, that was an expense that the firm was going to bear. Ultimately, again, it is all about pricing. In-house counsel do not come to outside counsel to buy hours.

It is not like buying hours at a daycare center or a ski lift. You are buying outcomes, or buying attempted outcomes. If you are going to buy attempted outcomes, you have a price in mind that you are willing to pay, and lawyers are going to try to figure out how close they can get to that price. Everybody is going to end up happy enough that it is going to keep going.

You’re not wrong. I wanted to grab something from your article. I know it is not your term, but there was a reference to the symbolic economy, or there is a prestige, like back to our pre-roll conversation. There is a perceived prestige that goes along with hourly billing. Let us talk about that for a second.

At a high level, and this also comes from my colleague, Mitt Regan, whom I mentioned earlier in his book on BigLaw, he and his co-author talk about how the billable hour serves two different purposes in the legal profession. One is what they call the material economy. That is the financial reward. You buy this, you get this in return. There is also an element of the billable hour that is related to what he calls the symbolic, or what they call the symbolic economy, which is the status and internal value.

It is why in the equation that I just gave, I defined compensation as the law firm’s compensation, because from the perspective of the client, that is who they’re paying. Anybody who has practiced in a law firm knows that the number of hours you work inside a firm, which may not be the number of hours that get billed to the client, is how you prove to the firm your worth to the firm. It is not necessarily just in a financial sense, but that is how you are valued in terms of promotion. That is how you are valued in terms of your bonus.

It is like a hitting average in baseball. It is a stat that shows how cool you are, how great you are.

Exactly. There is that piece of the symbolic economy. The other piece of the symbolic economy, and this is another reason why rates are going to continue to go up, is that counterintuitively, people think if they are paying more money, they are getting a better product. Actually, the incentive is not to fight on cost or on low cost because you will be perceived as not as good a lawyer. It is like, I think about it like when you go to a restaurant, and you are with other people, you never buy the cheapest bottle of wine on the menu.

Medium.

You go medium, because what does that signal? That signals “I am not cheap.” It signals “I care about quality.” Ultimately, there is a world, and I talk about this a very little bit in the article, but I think it is absolutely true, especially at top places, if you get too good a deal, you are going to go to a more expensive law firm.

I thought that was fascinating. This show is a big reason I do the show, which is for professional development for lawyers. That is a concept I am going to come back to because to me, it is feeding that ego. It is this perceived success. It is all these things that we as lawyers are sort of conditioned, if it is hard, if it is expensive, it must be good, and it is why I must want it in my life, and must want it in my career.

It is why the Wall Street Journal, I think it was just last week, published an article saying that there are some top lawyers in New York charging $3,400 an hour. We do not know. Maybe they are writing off half of that. What is the headline? The headline is the hourly rate because that is a prestige factor that has very little to do with the financial material reward.

AI’s Impact On Junior Lawyer Roles And Firm Leverage

I would love to talk more about that at some point. The last thing I want to discuss, just on the billing topic, is that I totally agree that law firms are built, and most professional services firms are built, on a pyramid model where your juniors and what you charge for your juniors are the economic engine that fuels the profits for those at the top, the equity owners, so to speak. One of the things I wonder, and you are teaching the next generation of juniors here, is there a concern that there are not going to be as many roles and as many opportunities for them to learn and bill? Do you see a funnel appearing?

If anybody who has listened so far can figure out that I am typically a tech optimist, I think tech, on balance, gets us more, not less. I am willing to wear that on my sleeve. This is the thing that keeps me up at night: how do we build infrastructure as a profession if we do not have a good way to incentivize those who are instructed to train the future professionals to build that profession? Again, not to sound like a broken record, this is not the first time we have had this discussion.

We have been having this discussion as long as technology has come around. I should not say if you remember, early on in this country’s history, the role of law clerk was mostly to copy down. There was basically a stenographer and a correspondent. That was the idea that you would sort of train as an apprentice by osmosis. Ultimately, we have been leveling up what the junior has been doing every single generation. The query is if we have hit that ceiling, and if you are a senior lawyer who does not need a junior to create leverage, can use technology to create leverage, where does the junior lawyer go?

Ultimately, I think that there are a couple of different ways this actually can work out fine for juniors. One way may be that juniors are not billed out to clients, which we know was starting to be true pre-AI. The firm was bearing a lot of the cost, and the firm was bearing it in part because they needed that junior labor, but it was also bearing that cost because that is the future of the firm. At least in a world where everybody gets a piece of the cake for the rest of their career, it makes sense to train people at your firm. They are going to find a way to do that.

The second thing is to say that maybe the senior lawyers are going to move even further up the chain. They can have even more cases, even more matters, and even more money. Basically, the first year is now going to play the role that a third year used to play, and a third year is going to play the role that a seventh year used to play. As a law professor who teaches skills-based classes, that petrifies me. How are you going to get them ready?

Students are not ready for that. I will tell you, legal education is not ready to train them to do that. There will be a messy period of time where, look, most of my life talking to lawyers is them telling me what we are not teaching them in law school. I have some strong views that are not my general good vibes-only approach on those topics. That delta is going to get larger before it gets smaller.

Ultimately, this is an opportunity to give junior lawyers greater exposure even earlier in their careers, because let us be real. Let us just be honest. Do you really need, as a junior lawyer, to spend a year formatting signature pages on a deal to learn how to do a deal? No. You need to be part of deals and watch deals. The actual task you were doing sort of justified you being on all the other emails. We can find other ways to justify you being on all the other emails. It is scary.

It is scary. Again, I think that the pyramid model, until it shifts, I think that there will continue to be a role for juniors. We can go round and round about whether you can survive a $3,400 or $5,000 an hour rate. We will never answer that question, but I do think just inherent. As you were prepping for this, I was thinking about how there are a number of positions that do not exist in law firms anymore. There is not necessarily a huge word processing department.

There are not even legal secretaries, who are really highly trained. People do not do dictation anymore. You have a redlining program as part of word processing, maybe there are no librarians, as there have been. There has been a winnowing of staff. As I think about it now, most of those functions were not billing you for. There is an inherent incentive to keep the juniors because they keep others fed. That is my final take on that.

There is some truth there. I also think there is a possibility that the same amount of legal work, and I actually fundamentally believe more legal work, will be done. There will be more hours to go around. New technology has shown us that every single time. Everybody said, “Now that we have these tools to do discovery, discovery will be faster.” Guess what? We then created a hundred thousand times the number of documents. It is far worse than it was, or better if you are charging by the hour to do the discovery.

More legal work will be done. There will be more hours to go around. New technology has shown us that every single time. Share on X

If you are one of the data companies.

There is a chance that you get a barbell effect where the big get bigger. The idea of being a junior lawyer who can go out on your own sooner and have a couple of key local clients and survive because all of the overhead and admin can now be done by a computer or by a software solution. I think there is a world where that starts happening more than we think it does, but that’s not an empirical prediction. That’s just the intuition.

You have got your ear to the ground. Just a teaser. I know you have another article coming out about AI and confidentiality. I am very fascinated by that. There was the recent decision out of the Southern District of New York, which was sort of like the first, I think, big splash around privilege. Just real quick, can you give us a couple of minutes on your views on AI and ACP?

AI And Attorney-Client Privilege Framework

This is something I have been thinking about for a long time, and I was writing about client confidentiality before generative AI became a thing, or at least a public thing. I have been thinking about it for a long time. It does not mean my views are any stronger than anybody else’s, but I do have strong views on it. Ultimately, the first thing we have to remember is that we often talk about the lawyer’s duty of confidentiality in the singular, but actually, we have a lot of different duties that come from a lot of different sources.

You mentioned attorney-client privilege. That is the one that is in the news right now because of the Heppner decision and then a decision out of the Eastern District of Michigan, Warner, which came to the opposite conclusion. You also have the model rules of professional conduct that have a huge impact. Over a dozen state bars have already passed formal opinions specifically about generative AI that almost all talk about some confidentiality. There are also confidentiality provisions written into attorney engagement agreements.

There are confidentiality provisions in court-ordered ESI protocols and everything else. The challenge, I think, for today’s lawyers vis-à-vis AI and confidentiality fundamentally is that every lawyer has a different set of responsibilities and a different set of risks that they need to worry about. There will be and cannot be a sort of checklist safe harbor. “If you do X, you are fine. If you do Y, you are not.”

My article fundamentally is saying, “Sorry, I wish I could give you sort of a checklist, but actually I think what we need is a framework.” Here are the kinds of things you need to think about. Here is how to think about them. One sort of teaser on how I think you should think about them is, what are the responsibilities you have been worried about all along with technology?

You have already answered that question. You know what your duties are to your clients. You know what is in your confidentiality agreements because you have email, you have cloud, and you have this, and you have that. That should inform your discussion of what your responsibilities are for AI. If you do not know what your responsibilities are, then you have no business creating an AI policy.

It’s because there is tech competency.

There is tech competency. That is one of the many responses, and that is the other half. Basically, I argue, you need to know what your responsibilities are. We need to have a real conversation about the risks. That is where, candidly, I think the New York decision got it wrong. I do not think the articulation of the specific risks is entirely accurate, even based on the face of what the tools do. Reasonable minds can differ. What responses do you adopt?

On the one extreme, you can just ban using the tools altogether. On the other extreme, you can allow any tool. In between those, there are lots of different choices. My argument again is that I am a history guy. Email is a great example. Originally, the first statement the ABA ever made that I know about email was “You cannot use it unless it’s encrypted.” 1986. The first two state bars to ask the question, “Can you use it?”

The first one says, “No, you cannot.” The second one says, “You can, but only if you get written informed consent from your clients. By 1999, the rule totally flips. The ABA puts out formal opinion 99-413, which says, “Actually, the default is you can use unencrypted email unless there are special circumstances or you need special protections.” Again, we are seeing this same layout in AI. The first answer is going to be fear, and we have to discuss what is real and what is not. The thing that I am not afraid of at all is that you are going to put information into an AI tool, even a free tool.

That is a spicy take, even a free tool. You put it in, and immediately, your opposing counsel is going to get that information out. I write many pages about why that happens almost never. Not never, but almost never. The actual risks that we should be primarily worried about are what happens when we give our whole document management systems to third-party companies that create an additional sort of access risk. That is very similar to your email, and similar to your Dropbox, and similar to everything else.

What was old is new again. My last question, hopefully not spicy. What is your pump-up song?

I love this question, by the way. I do not have quite such a great ending to my podcast questions. I will confess, I am not a huge music person, but I have two big pump-up songs. The first is Eye of the Tiger and the second is Pat Benatar’s Hit Me With Your Best Shot. I have been listening to those since I was a high school student, doing high school debate, when I got myself pumped up for high school debate. Nothing cooler than that.

You are not the first lawyer to choose those, so great, great classics. Jonah Perlin, check out his amazing podcast, How I Lawyer, and look for all of his hot takes in law review articles about technology and the law. Thanks so much, Jonah.

Thanks for having me, Stacy. This was fun.

  

  

Episode Resources

• Jonah Perlin on LinkedIn
• Jonah Perlin on X
• How I Lawyer Podcast
• How the Billable Hour Can Survive Generative AI,
• BigLaw
• Certified Corporate Governance Professional
• The New Rules of Executive Presence

 

   

Episode Resources

• Latitude Legal
• Leveraging Latitude
• C-Suite Series: What The CFO Wants In The Legal Department Kevin Rubin, CFO BetterUp
• What It Takes To Be A Good Lawyer In The Legal Department: Dana Lira Warner Bros. Discovery
• ResourcesPodcast Accelerating Business Velocity: Legal Ops in Focus with Mike Haven
• Using DEI To Level-Up The Legal Department: Valerie Portillo, Diversity Lab
• Don’t Say “No” Ask How: Developing A Team In The Legal Department Kate Dugan Energizer

About Jonah Perlin

Jonah Perlin is a Professor of Law, Legal Practice at the Georgetown University Law Center and Senior Fellow at Georgetown’s Center on Ethics and the Legal Professoin. In that role, he teaches and writes about legal communication, legal ethics, and legal technology.

His prior work is either forthcoming or has appeared in the Harvard Journal of Law & Technology, the Washington Law Review, the Yale Journal of Law & Technology, the Stetson Business Law Review, Legal Communication & Rhetoric, and the Georgetown Law Journal (student note).

Jonah is also the host of How I Lawyer, a Top-30 Careers Podcast for junior lawyers. Before returning to Georgetown where he completed his J.D., Jonah was a litigator at Williams & Connolly LLP and a law clerk to judges on the District Court for the District of Columbia and the Court of Appeals for the Second Circuit.