R&G Tech Studio Presents: Managing Principal and Global Head of Advanced E-Discovery and A.I. Strategy Shannon Capone Kirk

May 16, 2024
22:22 minutes

On this episode of the R&G Tech Studio podcast, managing principal and global head of advanced E-Discovery and A.I. strategy Shannon Capone Kirk sits down with data, privacy & cybersecurity partner Fran Faircloth to discuss how new and ever-evolving technology is impacting her clients, particularly generative AI, and the challenges that arise in litigation and compliance. She also discusses her team’s development of the AI Court Order Tracker, which examines standing orders and local rules on the use of AI in connection with court filings.


Fran Faircloth: Hi, I’m Fran Faircloth, a partner in the data, privacy & cybersecurity practice at Ropes & Gray—I’m located in our Washington, D.C. office. I want to welcome everyone to the latest episode of the R&G Tech Studio podcast. In this edition, we have my good friend and colleague Shannon Capone Kirk. Shannon serves as managing principal and global head of advanced e-discovery and A.I. strategy here at Ropes & Gray, and she is located in our Boston office. Welcome, Shannon.

Shannon Capone Kirk: Thank you, Fran. I’m really happy to be here.

Fran Faircloth: It’s so wonderful to have you. Before we get started and really jump into the content, can you tell us a little bit more about yourself and your background?

Shannon Capone Kirk: I have been a litigator for 25—I think going on 26—years now. Before being at Ropes & Gray I was a trial attorney at a mid-size boutique in Chicago. I’ve been here for going on 16 years now at Ropes, and then, the managing principal and global head of advanced e-discovery and now A.I. strategy. I focus mainly within the litigation and compliance realm but have impact and engagements with pretty much most all of the practice groups here at Ropes & Gray, given the pervasiveness of technology in the law.

Fran Faircloth: Yes, it sounds like something that all of our clients can use. Who are some clients you’ve been working with recently?

Shannon Capone Kirk: I touch on any form of corporation, mainly. There’s a few individuals in there, a few institutions, and educational organizations, but mainly, I would say, the majority are in the pharma or financial space.

Fran Faircloth: Got it. What are some of the main ways that the changing technology is impacting the clients that you work with, either for better or for worse?

Shannon Capone Kirk: There’s just such a multitude of ways, but where we’ve been focused on in the last year, and then, what I will be working on this year, is three main buckets:

  • One has to do with technology around mobile devices, and that meaning there is middleware or specialized apps to capture content on the vast multitude of mobile devices and app-based communications. There’s a lot of technology, nuance, and complication within that. We have gone very deep into all of it and advised mainly pharma and financial on those. And that is really impacting both of those industries greatly in the regulatory and litigation spaces because it all has to do with information governance, compliance with laws, and what is happening is this recognition that mobile devices are not devices. Yes, of course they’re literally devices, but the concern in the law is the data and the information: Is that really just a portal to a bunch of information? So, that’s having a great impact.
  • Another area, of course, everybody’s talking about is generative AI (“GenAI”) and what the impact that that will have on every single area of law. For my concerns, it’s counseling mostly around the litigation impact and compliance impact. A lot of what we do in my group is stabilizing everybody around the fact that we have been using a somewhat precursor to GenAI—although there’s so many big differences—that precursor being machine learning, predictive coding, etc., and how that actually is pretty established in litigation and court cases, etc. So, we’re moving through all of that with our clients and helping them to get aligned with what we’ve already done, and then, where we need to go in GenAI across the board.
  • And then, the last one—you might be surprised, but a big one for me, and I pretty much deal with this every single day now—is what I call distraction and misinformation around technology, and where we actually are with generative AI, and even still, machine learning. We see this in the media. We see it with entrepreneurs and vendors saying they have GenAI, and it’s viable. Then, when you start talking to them, you get them to admit it’s actually beta, so it’s level-setting on reality. What is actually really deployable and defensible now with GenAI for our clients? What are they already using, because they’ve got their own homegrown? And then, the important thing here is the precise language. Everybody—regulators, vendors, entrepreneurs, clients—should be using the correct language so that we’re regulating and strategizing all around the same thing, and we’re not talking in a way that is not matched because that feeds into this distraction, misinformation, and confusion.

Fran Faircloth: That makes so much sense. Can you talk a little bit more about some of the challenges that come up? For example, when people speak about “AI” just kind of blankly without specifying “generative AI”? We’ve had clients that say, “We want to use AI.” “We don’t want you to use AI.” Can you talk about the challenges of knowing exactly what everyone is talking about and how that emerges?

Shannon Capone Kirk: Absolutely. One challenge is when we get standing court orders from a variety of state and federal courts, and even some specialized courts now, they’re getting better and better, they’re getting more sophisticated, but we still have some out there that over-generalize that court’s order or requirements around “AI”—they define it in a way that could include machine learning, which we already have a body of case law on. Take, for example, some courts that might have this over-generalized treatment of AI and use over-generalized words around it, requiring certifications for the use of “AI” in court filings. Does that include machine learning, such as predictive coding, which is, for large law firms especially and government regulators, pretty much bread and butter at this point?

Another challenge is, for example, the SEC came out with a proposed rule last year surrounding what essentially was not phrased as GenAI or even AI, but for all intents and purposes, it was believed that that’s what they were going after. Their definition, however, was incredibly broad, and folks, including some folks within the SEC, criticized this proposed rule as including even Excel spreadsheets, so that’s a challenge.

Another challenge when we don’t use the precise language is an unrealistic expectation from clients right now, who are seeing everything in the media and hearing all of these hyper-inflated, aggressive marketing campaigns from vendors about GenAI—they’ve got GenAI, and it’ll cut costs, etc. It will someday for sure, and we embrace that, but some clients, I think, mis-perceive that it’s actually viable now and defensible now and we can use it now on large document reviews to cut down on costs. When you actually dig into it, right now, it is not necessarily deployable on a number of different kinds of document reviews that are costly. And when you do the cost comparison, the existing model of using machine learning—such as predictive coding, technology-assisted review, concept analytics—that actually winds up being less costly right now in our current state of affairs. I do see, though, the coming value of this for our clients. It’s just a reality set where we’ve got to get there—we’ve certainly got some ways to go for improvements on the existing tools for use in litigation and compliance.

Some of our clients are pretty sophisticated and have very sophisticated technology folks who have computer science and information governance master’s and doctorate degrees. They know how to design their own homegrown GenAI tools, and they have, which is fabulous, and they’re seeing some great efficiencies already internally. A challenge, then, is to encourage that, but at the same time, make sure that the legal and compliance folks are aware of the proliferating use cases internally in their own organizations, so that they are ahead of the game in litigation and compliance by way of preservation requirements, collection requirements, and advocacy around electronically stored information (“ESI”) protocols. So, all of the downstream ramifications of GenAI are the same downstream ramifications we’ve always had when major corporations implement pretty much any technology—it’s just faster now and proliferating in a wider way.

Fran Faircloth: That is definitely what I’ve been seeing as well. I know you’ve been doing so much to help clients stay on the forefront of these issues. Can you talk a little bit about the Court Order Tracker that you created to help clients sort this?

Shannon Capone Kirk: I love this thing because it’s colorful and fun, so, it looks pretty. But beyond that, I’m really proud of this thing. It’s a whole team effort—I’m one of the folks who is guiding, commenting, and asking for certain changes, but it’s a fairly big team effort of constantly crawling across a variety of sources to make sure we capture any state, federal, or specialized court standing order, or even specific judge order, around the issue of AI or GenAI. How we hope this helps clients is the following. There’s a multitude of ways that the courts are making requirements of litigants and parties on the use of GenAI or AI to the point that we’ve seen recent ones that require lead trial counsel to preserve all prompts and responses if used for litigation, and others that require counsel to certify if they have used AI in any portion of a filing. And there’s a question on that: Does that include research, or does that really only apply to the final court filing?

Fran Faircloth: Yes, wouldn’t most legal research involve some basic form of AI? Depends, again, like you said, on how you’re defining “AI,” right?

Shannon Capone Kirk: Exactly. And so, this is where we are trying to help clients with this tracker illuminate, “Is this court being fairly specific and sophisticated in its language?” If not, what’s the advocacy around that? Does that include this broad, broad, broad definition of “AI” to include legal research in preliminary drafting? If it does, what do we, as advocates, need to do to ask the court for clarification there? And how much is it that we really need to do in disclosures? So, if you don’t have comparisons to the other courts to figure out, “What are the wins telling us? What do we really think is going to be required here?” it’s really hard to do or to make that advocacy, which is another reason why we have the holistic view of the whole nation and mapping it out.

Fran Faircloth: That is so interesting. It seems like such a helpful tool. Is that something that anyone can find on the Ropes & Gray website?

Shannon Capone Kirk: They absolutely can: Artificial Intelligence Court Order Tracker

Fran Faircloth: Excellent. So, with this area changing so rapidly, if you look in your crystal ball to the next five or 10 years, what do you see as the main emerging issues?

Shannon Capone Kirk: I think that in about a year, maybe even six months, we will start to layer in GenAI into the whole document review process, whether it’s for investigations or actual document productions. And when I say, “layer in,” I mean it will become a tool that we use within the other tools we already use for those types of document reviews, such as predictive coding and analytics. I say that because we’re fairly certain that generative AI works on, right now, a more limited set of documents and could do a more limited set of tasks that we as humans can better validate and get comfort with—so that, I think, certainly happens within the next year.

Within probably the next two years, maybe, we will start to have wider use of generative AI in those same reviews, such that we start to reduce the use of the existing use of machine learning. That and what we’re already thinking about is the following: That is going to absolutely change staffing, project management, and budgeting. It will also require education of our juniors in two ways: Educating them on how to be good prompters of a GenAI tool in a way that we can validate to a court if we had to, and then, the other way—which is a concern of mine and a challenge that I want to get our arms around—if we are not having juniors do what they’ve traditionally done, which is document review, which is highly educational of the overall litigation process, how are we educating them on just strategy? What’s strong? What’s weak? What gets used? What doesn’t? How does it get used? How do you frame an argument around documents, etc.? All of that requires us to rethink how we educate our juniors. We also, because of those changes, will need to think about the challenges around ESI protocols and how we negotiate with the government and with opposing counsel on what, at all, if anything, has to be disclosed, and when, with respect to GenAI.

I also think that coming soon in the next couple of years—and I hope this isn’t just more of me just wishing—in mobile technology, that is a really big challenge for a lot of our clients, that the whole information governance and capture of mobile communications, right now, it’s clunky and difficult. There’s middleware out there, there’s specialized apps, but there’s difficulty in capturing iMessage. There’s difficulty in a multitude of ways, it’s actually even hard to chart out, but I think and hope that within the next few years, somebody’s going to come up with a solution that makes it better streamlined and easier integration for large organizations to have better control around mobile communications.

Fran Faircloth: That seems like a really good opportunity for technology to step in and fill the gap there. This is really interesting. I’m sure we could talk a lot more about all of the things we see coming in the next few years, but I want to make sure we leave a few minutes for our listeners to get to know you a little bit better. We left a few questions here at the end which actually don’t have anything to do with the law or technology, necessarily. So, it’s pretty quick, rapid fire, a bit of a lightning round here at the end. First, can you just tell us a little bit about yourself, your family, and where you live?

Shannon Capone Kirk: I live in Manchester-by-the-Sea, which is absolutely nothing like the movie—it was a pretty dark movie, really well done, but that’s where I live. I have one son who is in college. And I live there in Manchester with my husband and my three kitty-cats. I am also a published author.

Fran Faircloth: Can you tell us a little bit about your books?

Shannon Capone Kirk: Sure. I have a secret mission in life—I’m a thriller writer. I have a psychological thriller. I have a couple of Gothic thriller/horrors. I have a legal thriller. And now, I’m in the process of trying to sell to a publisher a sci-fi thriller. So, this is probably the first time I’m disclosing publicly that I actually have a personal goal to publish every single style of thriller.

Fran Faircloth: That is awesome. I think I’ve told you I have a book that I’ve been wanting to write, and it is a bit of a thriller, but can never seem to get past the first couple chapters. So, maybe I’ll have to workshop that with you.

Shannon Capone Kirk: Yes, Fran—and I very much highly encourage you to do it, because I know whatever you write is going to be brilliant.

Fran Faircloth: We’ll see. I don’t know how you find the time to do it—that is my challenge. I know you also, like me, read a lot. Can you tell me about one of your favorite books or one of the best books you’ve read recently?

Shannon Capone Kirk: Yes, so I keep trying to find a book to knock this one off the top pedestal, and I just can’t find it, but it’s The Incarnations by Susan Barker. I did a whole interview on this once. It’s insane how much I love this book. I think I’ve read it three or four times at this point. It’s so good. It is so incredibly unique. It’s basically this Beijing cab driver who starts getting all of these mysterious letters from his “soul mate,” and the soul mate claims that they have been together for thousands of years. So, then interspersed between modern-day Beijing, when he gets these letters, you get these retrospective chapters where they’re together thousands of years ago. It’s so incredibly vivid, and at turns darkly comic, at turns horrific, and at turns historic. It’s so different from anything I have ever read, and it’s addicting.

Fran Faircloth: I will definitely have to add that to my list. I have not read that, so it’s exciting to hear about a new book that I haven’t even heard of before—thank you for that. To close us out, if you could imagine your perfect day or happiest dream, what would that day look like?

Shannon Capone Kirk: It is some variation of me being in a perfect solitude: writing. And that’s either in a really nice comfy, cozy cabin on the ocean, where I get to hear the waves, or I’m in Vermont and I can hear the wind and there’s a river, and I can go fly fishing in between breaks. But it’s some variation of that—that, to me, is a heavenly dream.

Fran Faircloth: With those peaceful thoughts, I want to thank you, Shannon, for joining us today. And thank you to all of our listeners. This has been the R&G Tech Studio podcast. It is available on the Ropes & Gray website, on the R&G Tech Studio podcast page, and also wherever you get your podcasts. Thank you, everyone, for listening.

Subscribe to R&G Tech Studio Podcast