“Hearsay Culture” Podcast Founder David Levine Discusses New Research, AI And More
David S. Levine is a professor at Elon University School of Law and the founder and creator of the award-winning Hearsay Culture podcast and co-founder of the new Hearsay Culture Network (found at hearsayculture.com). As a radio and podcast host, he has conducted over 300 interviews. In November of 2023, he was selected as a fellow at the University of Milan’s Information Society Law Center, where he will conduct research and collaborate with international scholars over the next two years. The law center is a multidisciplinary hub providing researchers opportunities to explore issues related to the relationship between law and digital society.
Levine, who holds a J.D. from Case Western Reserve University School of Law and a bachelor’s degree in industrial and labor relations from Cornell University, was named the Jennings Professor and Emerging Scholar at Elon University School of Law for 2017-19. He has been credited with creating the scholarly field of trade secrecy and information access and has published many articles, including “Generative Artificial Intelligence and Trade Secrecy,” which appeared in the Journal of Free Speech Law in 2023. In 2019, he co-authored a textbook titled “Information Law, Governance, and Cybersecurity.”
We interviewed Levine to discuss his current research projects, artificial intelligence and the law, his path to becoming a law professor, how he began the Hearsay Culture podcast, and more.
In the fall of 2023, you were selected as a fellow with the University of Milan’s Information Society Law Center. Could you tell us about the research or project you are working on as part of the fellowship?
I was very excited to join Milan’s Information Society Law Center. I applied with the view towards working with them on generative AI regulation. The EU has taken a lead in that space, and the European regulatory framework is quite different in a lot of ways than in the U.S. So I was looking for opportunities to collaborate.
What I’m working on as part of that fellowship, are the broad questions associated with information access and generative AI. I’ve also been thinking a lot about whether there is a holistic theory of information law that we could apply to everything from generative AI to contracts that were written in the 16th century on parchment and sealed with wax.
Would you say that that’s something new in your research – looking back to these older contracts?
It is new in the sense that my scholarship has generally focused on what’s happening now. I made the transition into the academic world from practice in the mid-2000s. That’s when the internet was really starting to explode. Logic dictated that if you wanted to be doing impactful, cutting-edge scholarship and having the opportunity to do it, that was the area. Generative AI raises even more fundamental questions in some ways than the internet did, because it competes directly with artists and authors in a way that the internet didn’t. It gives an opportunity to view information law historically and holistically.
How has it enriched your research to be able to collaborate with international scholars?
I’ve done that informally for my entire academic career. The reasons for doing that are a few. If you’re talking to someone who is in France, for example, about freedom of speech, there are strict rules with regard to using Nazi images on social media for obvious reasons in France that, up until recently, really didn’t exist in the U.S. There are different conceptions of freedom of speech. In other words, law and values really is one reason, but also there are different goals for policy.
If I’m talking, as I have in recent years, to scholars based in Africa about access to medicine, it’s a very different conversation when the manufacturing capacity doesn’t exist or is limited on the entire continent. Most of the scholars that focused on Africa during the Covid pandemic understandably talked about how important it was that the manufacturing capacity exists. If I talk to scholars in Europe, particularly Western Europe, it can be a very different conversation. It might be about effectiveness first or cost. These are not necessarily mutually exclusive, but they’re different policy priorities. So being able to talk to people who have different perspectives with regard to outcomes is important.
And then of course, in my areas, there are different experiences with technology. It’s very important as a scholar to be able to step out of your own frame and understand these issues from the perspectives of others, particularly as it relates to technologies which obviously span the globe.
I’ll give you a great example. During the negotiations of the Trans-Pacific Partnership Agreement (TPP) in the 2010s, I advocated for an open negotiation process, and it would boil down to access to the texts that were being discussed, and for a variety of reasons, those texts weren’t made public, but occasionally they were leaked. Once they were leaked, the dozen or so countries that were involved in negotiations had stakeholder listening sessions where they would allow the public to apply and speak for eight to 10 minutes to negotiators for the countries involved.
Because of these opportunities, with the support of Elon, I flew to Auckland, New Zealand, and a few other places. I spent 36 hours in Auckland. I did my talk about access to information and the ability to allow stakeholders that weren’t part of the negotiating process to have input into the negotiations, because they would be impacted by the outcomes. I came out of the room, and there were three representatives of the indigenous Māori tribes standing in the hallway. They came up to me and basically said, professor, thank you for speaking for our interests, for making the case for access.
To say I was moved was an understatement because this is an enormous privilege of being a tenured professor.
To be an advocate for interests that are not well represented and that may not have the political, societal, or economic power to be heard is one of the things that tenured professors can do. I consider it an honor.
In 2023, you wrote an article on artificial intelligence and trade secrecy. In it, you analyze ChatGPT’s ability to disclose trade secrets, which occurs in some cases by users who input information through answers to prompts. Could you describe some of the risks involved with this new technology? Is there a way for trade secrets to be protected in an age where AI is easily and readily accessible?
I wrote this article for policymakers. Generative AI is a technology that, through its probabilistic language matching and thinking, can predict what the next word should be and create information. And it creates information in a way that human beings may not do because we don’t operate the same way, and that’s very exciting because we can get new amalgamations and kinds of information as a result.
At the same time, through the massive training sets that ChatGPT and other large language models use, there are risks to maintaining trade secrets. As a general matter, when you put information that you consider to be a trade secret out of your control, you put it at risk. There are ways to deal with that in the world, including what the law calls “reasonable efforts” like non-disclosure agreements, but obviously you’re not entering into non-disclosure agreements with OpenAI when you share it willingly without such an understanding. As I wrote about in the article, even early on, within months of ChatGPT being publicly available, large companies found that their trade secrets were being spit back in queries. So, that was the focus of this article.
Do I blame employees for putting trade secrets into ChatGPT? I mean, as a lawyer, I can say, bad idea. And as an employee, assuming you’re busy – assuming you don’t have all the time in the world, or even enough time, which many employees don’t in our fast-paced technology environment, or you know your competitors are using it so you feel you have to also – its no surprise that its being used, and likely more than most are willing to admit.
The next piece I’m working on, which relates to what I’m doing with Milan, is an interesting question of information access. With every technology that I’ve written about, you can make a very strong argument that if the code is available to researchers, to governments, to civil society groups and others, they could learn something important from it. It turns out that generative AI, or AI more specifically, is different. And how do we know that? It’s because the people inside the companies themselves are saying that they don’t understand how many decisions are made by the generative AI. If that is true, then it raises the question of what benefits derive from putting the code out there, particularly if there’s costs associated with deterring innovation?
This is the first code-based technology in my career, particularly involving computers, where access to the code may not provide the same benefits that, for example, it would with voting machines, or it would with Breathalyzer machines, and other kinds of code-based technology. I find that very interesting and challenging in terms of the benefits of transparency versus the benefits of secrecy.
You have been credited with creating the scholarly field of trade secrecy and information access. How did your interest in this area of law come about?
I’ve always been interested in why and how people kept secrets. I am an amateur historian of World War II and have been from a young age, which quickly got me reading about espionage and counterespionage.
There was a terrific book written about D-Day called “Bodyguard of Lies” by Anthony Cave Brown, and it was about how the Allies deceived the Germans into thinking that D-Day wasn’t going to happen where and how it did. It just fascinated me. Even from a very early age, I was interested in secrets – in why secrets are kept and how people use secrecy, and conversely, transparency to achieve outcomes. Fast forward to when I wound up at Stanford as a fellow. In my first article, I described how a hacker had discovered a vulnerability in a Cisco router. My reaction was, Cisco must be pretty happy that this hacker found this vulnerability that they could now fix. And I was wrong.
Not only were they not happy about it, but they were alleging trade secret misappropriation. And I thought, wow, that’s not the main reaction that I expected.
This is 2005 and fresh on the heels of Bush v. Gore. It quickly got me looking at voting machines and I found that voting machine code is a trade secret that was inaccessible, even to government regulators, and the issue hadn’t really been studied. So my first article on secrecy and unaccountability, which was published in the Florida Law Review, is credited for being the first article to fully explore the issue. As a scholar, to find an area, no matter what it is, that has not been written about heavily is like striking gold. It’s very difficult to do.
In that sense, for whatever reason, I was fortunate that it hadn’t been covered. My broad concerns here involve accountability, fairness and access to information – trying to have some particularity where it’s beneficial to society to share trade secrets and to level the playing field in terms of access, so that we can make better decisions. On the other hand, keeping trade secrets is very valuable and beneficial to society in a lot of formats. That makes it a very rich area to study because it’s not a one-size-fits-all area.
Did this experience lead to more research beyond the first piece because you enjoyed this area and wanted to share your work with others?
Yes. The first article focused on voting machines and other technology used in public infrastructure, and it quickly expanded into other areas where trade secrecy plays an outsized role in keeping information concealed, like breathalyzers and hydraulic fracturing. More recently, it involved Covid vaccine manufacturing processes.
It’s important to note that trade secrecy is not well understood. It’s widely understudied in the scholarly arena. It does not get the attention it deserves. I have, for years, referred to trade secrecy jokingly as “the fourth of three intellectual property (IP) regimes.” What I mean is, if you take an IP survey course in most law schools, the focus is going to be copyrights, trademarks and patents. Maybe trade secrets will get mentioned in one or two classes out of 30 classes. In other courses it may not be mentioned at all, even though it is a gigantic part of the innovation ecosystem, and even though there is a wide range of its applications.
Before joining Elon Law, you practiced law in New York. Could you tell us about moving from private practice to teaching law?
I never thought I would be able to be a law professor because academic hiring is heavily credential-driven, meaning that the majority of people who get hired for entry-level doctrinal positions at ABA accredited law schools went to about a dozen top-ranked law schools, of which Harvard, Yale, and Stanford dominate.
As a graduate of Case Western School of Law, I never really thought I would be able to be a law professor, and more or less put it out of my mind. But then 9/11 happened and I (like many) quickly understood that you only live once, and when opportunities arise, you need to go for them because you don’t know if they will ever come around again. I already knew that I didn’t want to be a law firm partner. I really enjoyed mentoring, and before law school, I had worked in government for a year for my state assembly member. I always had interest in public policy and government, and I just didn’t really aspire to be a law firm partner.
So in earnest after 9/11, I started looking at things that I could do outside of day-to-day practice of law. I looked at potentially getting a Ph.D. in history but decided I didn’t want to do another five years of school. Then I learned about these things called fellowships. I thought ‘maybe I could become a law professor by first being a fellow?’
Long story short, Professor Larry Lessig and Jennifer Granick hired me at Stanford Law School in 2005, which blew my mind. I couldn’t believe it. My wife Heidi, a public school teacher, deserves all the credit for encouraging me to apply to Stanford (and has since taught me more about teaching than anyone I know). As you might imagine, my response to her was ‘why on earth would they hire me?’ But it turned out that Lessig and Granick wanted someone who was a litigator because Lessig, at the time, was working on issues involving copyright law on the internet. And I had done work for entertainment industry clients in copyright and trademark litigation. I had a lot of experience as a litigator, and after a few interviews, they hired me.
I regularly talk with students about careers, and I always tell them, as the old saying goes, if you don’t try, I can guarantee the outcome. I would have bet a lot of money at that time that I’d never get hired by Stanford. So if you have a chance to chase your dream and you’ve done the work, if at all possible, go for it. You only live once.
I spent two years at Stanford. I would not be a law professor if they had not hired me and given me the opportunity to try to transition from practice to full-time doctrinal teaching. I am forever grateful for that. So I did public interest litigation there and wrote that first article. That fellowship opened the door for me to be a law professor.
Why did you choose to become a law professor?
First, I love mentoring. In practice, my view was people are going to do a better job if they understand why they’re doing it, but on a personal level, I enjoyed it. I also very much enjoyed debating issues. In fact, what I enjoyed most about practice was debating and arguing the law on behalf of clients. There were things that I didn’t enjoy, but I loved finding and briefing legal arguments and thinking through and reaching the right outcome.
Ultimately, I realized that I really enjoy helping good people have professional success. By helping a paralegal figure out how to do a task better or having a first- or second-year associate be able to learn from my experience, felt very rewarding. I realized all of this is about teaching.
We should all be blessed with great teachers, and in that regard, I’ve been lucky. Perhaps most influential, I had a middle school social studies teacher named Carl Oechsner. Carl was a legendary teacher in the Ossining, New York, public schools. He had a classroom that was covered literally from the base of the walls and on the ceiling with newspaper clippings and old bottles and everything else – total immersion in history and the social sciences. Carl oversaw the “Student Rights and Responsibilities Committee,” where he would have eighth graders go around our middle school and talk about student rights to speech, privacy, and so on. He had a Supreme Court decision eighth-grade lunchtime study session – he’s not a lawyer – but he summarized major Supreme Court opinions in writing for us (way before ChatGPT!) and we talked about them. I had teachers like that; I didn’t grow up around lawyers so it was Carl who first mentioned that I might be interested in law.
I had teachers as an undergraduate and graduate student, George Boyer at Cornell, Kevin McMunigal at Case Western, and others, who had an enormous influence on me because they gave of their time, and they mentored me. I also have colleagues, particularly Steve Friedland, Eric Goldman, and Howard Katz, among others, who have been incredibly generous in helping me learn the art and skill of teaching law. Whatever I do now that’s good in the classroom and professionally, I model after and owe it to them and my wife Heidi. I was also fortunate in law school to be a summer clerk for Bankruptcy Judge Adlai S. Hardin (SDNY) and to work with lawyers like Jennifer Granick at Stanford and Bob Mercurio of Windels Marx, who demonstrated how to apply and practice law with integrity and compassion.
I wanted to have the opportunity to have that kind of impact on junior colleagues and students, which is, of course, the greatest reward of teaching and mentoring.
There is something special in looking back, as you are pouring knowledge into your students, and thinking about the long line of mentors who spent time mentoring you. That must be rewarding.
Oh, absolutely. There is no one who has had success professionally, or maybe even personally, who can’t point to at least one person like that, right? The tragedy is that not everyone is that fortunate.
I encourage my students to find someone for professional advice and guidance, in the law school or elsewhere. This is how success happens. None of us are self-made. Anyone who has had any success has been fortunate to have people around them that offer wisdom and encouragement.
The tragedy is that there are many people who are not as fortunate because of life circumstances, luck, timing, and resources. One of my goals in my own small way is to try to level that playing field as much as I can, and help my students find professional success and happiness with integrity.
In 2019, you co-authored “Information Law, Governance, and Cybersecurity” with Sharon K. Sandeen. You have described how you wanted to create a textbook that would, as you say, “teach new lawyers and young lawyers how to think broadly about how information is collected, accessed, regulated and used.” Could you talk about why it is important, especially now, for new and young lawyers to think broadly in this way?
We don’t think of information law as a field. We tend to think, well, there’s information everywhere, and that’s the world. But as it turns out, thinking about, using, and managing information is a big part of what we do as lawyers. And it’s a big part of what we do in society.
Sharon Sandeen and I collaborated on putting this book together, and the book is structured around bodies of law, like contracts and privacy, and also information governance, how you manage information inside organizations, how you manage information inside governments, national security and so forth.
Because of the nature of the threats to information access, particularly in 2019 and still today, we discussed cybersecurity. This is a book that has very few precedents. There really has not been an information law textbook quite this way, and it’s even more timely now that generative AI is in the world. I’m very proud of what Sharon and I put together.
As the host of the Hearsay Culture podcast, you have produced hundreds of interviews – over 300 – since 2006. What led you to create the radio show?
Hearsay Culture is on Stanford’s radio station, KZSU-FM (Stanford University) and airs right now on Saturdays at 7 p.m. Pacific/10 p.m. Eastern. I did radio in high school. Ossining High School had an FCC licensed radio station called WOSS-FM. I became the station manager as a senior, and I loved it.
I did a show called “Inside the NHL, (the National Hockey League”), with my friend Paul Caley. We talked about hockey for an hour. Paul and I had no “inside access” to the NHL whatsoever, of course. We occasionally would get a listener phone call, but basically, we were just two high school nerds talking about professional hockey as best we could. And I also DJed.
When I got to Stanford, I discovered that there was a radio station at Stanford called KZSU-FM that was located across campus from the law school. I thought I would love to get back into radio, and I could do an interview show because I have these amazing potential guests right at Stanford. And no one’s doing this. So, I walked across campus to KZSU. When I told the program director my idea and that I had done radio in high school, she fast-tracked it, and I was on the air soon thereafter.
My goal was and remains to bring good work to a wider audience. I started focusing on not just “big names” or people at high-profile schools like Stanford but finding academics at places that may not have the same ability to get attention. This was in 2006, just as social media was starting to take off. I brought guests on the show to talk about their articles or books or their ideas in a long-form, serious, but engaging and fun style. I read every article or book about which I’m interviewing, so was usually quite happy with the questions that resulted. The show was recognized by the ABA as a Top 100 Blawg of 2008, and it built a dedicated audience. I’m back to it now by launching the Hearsay Culture Network to an even wider audience and broader scope.
How has the show evolved over the last few years into the Hearsay Culture Network?
I did the show regularly for about 10 years from 2006 to 2016. Then, starting in earnest right before the Covid pandemic, I started working on relaunching it with Howard Greenfield as my partner. In late 2023, it relaunched as the Hearsay Culture Network at hearsayculture.com.
We’re doing new programs and more shows with my friend Denise Howell, an experienced podcast host who did a show for years called “This Week in Law” for Leo Laporte’s This Week in Technology network. She and I are doing a show together called R&D with D&D. Denise is doing her own show called Uneven Distribution. I’m doing my KZSU show, and we have other shows and content in the works for subscribers. We have already posted a bunch of shows, and you’re going see a lot of content going forward.
Hearsay Culture has always significantly informed my teaching and research, and even more so now that we’re expanding coverage to include not just primarily law and culture, but also business, entertainment, and technology. I’m very excited about its future and how I can bring what I learn to the classroom.
Describe your approach to teaching the law and why teaching is important to you.
The big picture is that the law is the bulwark against chaos in our world. I’ve been saying that for years. We are in a profession that, in our best incarnation, is structured to focus on facts and evidence in order to form opinions. We do it imperfectly, and it doesn’t mean that all biases and human failures are eliminated, but as a profession, we adhere to the concept that facts, evidence, and law drive outcomes, not the other way around, even if that doesn’t appear to always be the case. Along with that is understanding why we have law in the first place, the theory that underlies it and the outcomes that result. That’s my focus in all my classes.
I also feel very strongly that it’s important to keep classes engaging as part of the learning process. I go out of my way to try to make material as interesting as I can, and I do it through interaction, listening to my students, and occasionally, humor, such as it is. Also, I do a good amount of cold calling. I use online forums where people can write answers to different questions. And widely available office hours.
This past year, I started doing oral exams, which are not common in law schools. I sat students down both in my IP class in the fall and in my Contracts class one-on-one this past winter for 10 minutes, with the questions that they had in advance, and had a conversation about the law. I decided to do this, in part, because the technology that we use, I think in a lot of ways, has diminished the skill of discussion of law orally. We have a generation of people coming up in law schools who are, in a lot of ways, much more accustomed to texting than they are speaking.
How do you talk with your students about AI?
My approach to teaching is very much driven toward making sure that students understand the basic rudimentary skills associated with being a lawyer. That’s number one. With regard to AI in particular, I have said to my students that whether we like it or not, this technology is here, and so we have to learn what it can do and what it’s capable of doing. I don’t have a crystal ball to say whether it’s going to flood the profession in the next year, or five years or 10 years. But it’s obviously here.
Generative AI has enormous potential for good, but I try to be candid and give direct advice for where we are now. As a result, I’ve said to my students that if you want to render yourself obsolete, do things like not brief your cases. Briefing cases and developing reading comprehension are the skills you need as the foundation for everything else that lawyers provide. ChatGPT is very good at summarizing facts, and if you can’t summarize facts yourself, you can be sure that this technology will do it for you. But unless you can do the work that ChatGPT and the other technologies already do, you won’t be able to develop the skills that generative AI is not great at doing, like understanding nuance, divergent perspectives, and counterarguments, addressing new scenarios that can’t be analyzed through the training data, and providing context, as a few examples. The result is you may be out of a job as a lawyer – if not today, then in the foreseeable future.
If it were up to me, ChatGPT would not have been released unilaterally by OpenAI’s Sam Altman to the entire world on November 30, 2022. We would have spent a lot more time as a society not only recovering from Covid, but also understanding generative AI’s capabilities and limitations, its risks and benefits. But we don’t live in Levine’s Magic World.
Instead I’ve taken what I think is the reasonable approach, which is, focus right now on developing your skills that you always needed as a lawyer, and at the same time we can explore, because we are really exploring together, what the technology is capable of doing and what it is incapable of doing very well.
It’s an incredibly exciting, if unnerving and uncertain, time to be training future attorneys. Without respect for the rule of law and the lawyers and judges who apply it, our country is done. It’s a privilege that I cherish, and I’m grateful to be able to do my part to help teach future lawyers and maintain and improve the profession that I love.
Jessica Junqueira is communications manager for the North Carolina Bar Association.