Listen on Apple Podcasts, Spotify and YouTube.
Welcome to AccelPro IP Law, where we provide expert interviews and coaching to accelerate your professional development. Today we’re featuring a conversation with Mark McKenna, faculty Co-Director of the UCLA Institute for Technology, Law & Policy.
Artificial intelligence is transforming academia, business and entertainment. The rise of generative tools such as ChatGPT and Midjourney are leading to new questions surrounding copyright and artificial intelligence. Precedent for generative AI and intellectual property law is being established on a regular basis.
Mark McKenna is Professor of Law and Faculty Co-Director of the UCLA Institute for Technology, Law & Policy. In this interview, we discuss generative AI and IP law, trademark protection and more.
Listen on Apple Podcasts, Spotify and YouTube.
Interview References:
Mark McKenna’s Lex Lumina and UCLA Law profiles.
4:06 | Aunt Jemima Mills Co. v. Rigney & Co., 247 F. 407, No. 42. United States Court of Appeals, Second Circuit. (1917).
5:44 | Rogers v. Grimaldi, 875 F.2d 994. United States Court of Appeals, Second Circuit. (1989).
13:38 | Google Library Books Project. (2012).
13:47 | Prima Facie Copyright Infringement. (2011). The IP Law Blog.
Supplemental Materials:
Authorship Issues with Generative AI. Harvard Business Review.
Minerva Surgical, Inc. v. Hologic, Inc., No. 20-440. On Writ of Certiorari to the United States Court of Appeals for the Federal Circuit. (2021).
TRANSCRIPT
I. GENERATIVE AI & IP LAW
Neal Ungerleider, Host: What is the Institute for Technology, Law and Policy?
Mark McKenna: The Institute for Technology, Law and Policy at UCLA is formally a joint venture between the law school and the school of engineering. It's meant to be an interdisciplinary program that draws folks–especially from those two areas–but hopefully much more broadly than that around campus. We want to bring people with lots of different disciplinary backgrounds into conversation about important and especially emerging questions about technology, policy and regulation.
The basic recognition is that all really important and hard technology problems have dimensions that lie outside the expertise of any particular set of people. Lawyers are obviously important in designing policy, but lawyers have to understand the technology. And the technologists obviously have a significant role both in sort of the way they design technology in the first instance, but also in teaching people how it works so that we can make the right kind of policy.
I know there obviously are lots of roles for folks trained in a bunch of social science disciplines, and even in the humanities. The Institute's only a couple years old, but we feel like we've made a lot of progress.
NU: Can you give an example of some of the technological problems you mentioned?
MM: The one that's getting probably the biggest amount of attention this year are questions surrounding generative AI and thinking about ChatGPT and some of the image technologies, and thinking about what problems that might pose. Even within my own field of intellectual property, there's hot debate about what authorship means in an era of AI creation of new information and new imagery.
What happens when the AI tools are being trained on massive data sets that include lots of other people's own creative works?
We are also dealing with all kinds of other questions about regulation of AI—for example, thinking about algorithmic discrimination. And the sort of advanced version of that is thinking about governance of algorithms and governance of emerging technologies in a variety of different spaces.
Our institute this year did a whole bunch of programming around threats to journalism and the relationship that that had with technology both in threats to the business model, in terms of the way that technology is sometimes used for surveillance purposes and as tools to suppress the ability to gather and disseminate information.
So, you know, it's a pretty wide range of different things, but technology touches kind of everything now so we have to be wide.
My first teaching job was at St. Louis University, and I was lucky in the sense that it was just after or in the middle of the “dot com” era. And a lot of law schools were realizing that they needed some expertise in this area, and they didn't have much. And so there was a supply and demand imbalance that was in my favor. That's how I got into the field. I feel very lucky to have gotten into it, especially when I did. The thing that was the attraction for me, was always that it was a field that was all about all the things that were developing and changing in the world, that people would see in terms of new technologies.
I was going to law school when the commercial internet was being built, and that meant a whole constant set of questions about whether that was a revolutionary technology that meant you had to start from scratch and rethink all the legal rules. Or whether instead it was sufficiently analogous to other things that you could tweak. And that's a recurrent issue in all technology laws. When are we at these points where the technology, it's sufficiently different, and we need really new thinking about things.
—
II. TRADEMARK PROTECTION & FREE SPEECH
NU: How do you reconcile trademark protection with things like free speech competition values? From your perspective, how does that play out?
MM: Our view has been pretty consistently that trademark law used to be much narrower than it is now. It used to be, basically, it gave you trademark rights that you could enforce against other parties who were making more or less the same kinds of things.
Maybe not directly competing, but very similar sorts of goods. The concern was that consumers would think those products came from the same company. So, if you went way back to one of the cases I teach my students - and this case is a bit of an anachronism now - but if you went way back, court’s worried about one company's using Aunt Jemima for pancake mix; the other one using it for syrup.
Well those don't directly compete with each other. But because they're complimentary products, you could easily imagine them coming from the same place.
So they developed the likelihood of confusion factors that every circuit has, and those factors are meant for cases like that. It's trying to figure out when it is considered infringing, even though you don't have directly competing goods, because the products are sufficiently close to each other.
Over the course of the 20th century, courts took a much broader view of what it meant to say that the goods were related. And they especially started developing this idea of what they called sponsorship, or affiliation, confusion. And that really slung the doors wide open to claims that companies have made against people producing movies.
So Wham-O sued the producers of a movie because they used a Slip ‘N Slide in the movie. And the Slip ‘N Slide, they didn't hose it down. And so when the people dove on it, they didn't slide. They hurt themselves - a joke in the movie. And they sued, and their argument was, “People will think that they got our permission to use this in the movie.”
The luxury brand companies are suing like this all the time. Louis Vuitton sued the producers of the movie The Hangover Part Two because they made a joke about Louis Vuitton in that movie.
I think a lot of consumers have been convinced that anytime you make a reference to a brand owner that you must be getting their permission. And since sponsorship or affiliation doesn't really have any precise content to it, these claims are often plausible. So, and my view has always been that Rogers versus Grimaldi - which is a very speech protective rule - that rule basically says if you're using a mark in expressive work, it's not considered infringing.
It's insulated from infringement liability. As long as the use is artistically relevant, meaning it has some relationship to the work. And courts have interpreted that requirement to be very low, right? Because they don't want to get into the business of trying to decide which things count as art or not.
So basically a very low threshold, as long as there's some artistic relevance. The language from the Rogers case is actually as long as there's any artistic relevance whatsoever. And then also, as long as it's not explicitly misleading. So explicitly misleading is not supposed to just be, people might wonder.
It's supposed to be, no, you said something explicit. Like, this is the officially licensed version of, or something like that. And so every circuit now that has considered the question has adopted some form of that test. The Ninth Circuit is a very strong version. The Second Circuit's a little less good in that respect, but the reason they developed that doctrine is because these claims became plausible under a very expansive view of trademark law.
And so, what I think has happened over the last few decades in trademark law is that because the internal boundaries have gotten so broad, courts have had to develop a bunch of prophylactic doctrines to say, “Well, you can't make these kinds of claims here.” My view is Rogers is really essential in that respect. Because otherwise, every time you make a movie and you make a reference to a brand, you're gonna have to get a survey and litigate a case to decide if you can do that. No surprise, the Motion Picture Association strongly agreed with us about that because they said, “This is gonna be a real problem.” So, I think the answer is that trademark law is supposed to apply only to commercial speech. Trademark is ordinary commercial or commerce clause regulation.
It doesn't have any special status in the Constitution, so it has to give way to speech. There are some hard questions, obviously about how you decide which things are commercial speech versus non-commercial speech? But it clearly has to be the case that there are limits that are imposed by the First Amendment, because they're not being imposed internally to trademark.
NU: Does the fact that the internet democratized production making these cases more challenging? Just as an example: being able to make a video on your computer. Instead of having to go out and get a camera; find an editing facility. Or being able to make physical objects just by going online; sending production plans. Has that made these cases more challenging, or is it more of a scale issue?
MM: I think both. So, I think there are some things that technology enables. There were always analogous versions of offline, right? Think about memes–memes as a concept, and as a way of creating culture. Those have existed forever. But obviously, the virality of social media and the ability to copy and produce those things has radically changed the scale of the distribution. So one component of that is to say: Scale changes the nature of the legal question. Because things that might have gone undetected or below the radar, just not really have caused a lot of harm, now are circulating at a level that is easier to spot. I'm not necessarily saying memes are an example of that, but certainly file sharing is an example of that.
You know, people were back in the day putting two boomboxes right next to each other, with a blank tape, and recording. That was happening all the time. But to do that you needed two things in physical proximity to each other, and you needed two tapes.
And every time you made a new copy of those things, the quality degraded, right? And so cloud sharing was just like that on steroids, but in a way that was much more disruptive to the market. So yeah, sometimes the creation of new technology takes things that have existed for a long time, but it really changes the scale of them. Which changes the sort of economics of what's going on.
I do think that there's no question though, that the ease of production now and the tools, means lots more people can be creators than ever was true before. It used to be that you needed studio time, you needed expensive equipment, right? You needed expertise. And you know, you're doing this podcast–I think probably sitting in your family room, right?
And my son, he's making music, and he's making it literally using his laptop computer at the public library with a couple of pieces of equipment that aren't super expensive. And then, he has some people who are helping him mix it. And so, for a very small amount of money relative to what it would've cost before, they're making studio quality music.
So, that's the sort of great side of technology, right? It enables lots of creative activity. But it still requires lots of questions about how that stuff gets distributed and what role is there for platforms and that kind of thing? So technology has a sort of duality to it a lot of times, right?
You might even say, right now we're living in sort of a golden age of creativity. Because it's made possible in so many forms, including the ability to remix stuff. But, that also, from the perspective of some people, means more uncompensated usage of stuff that gets sucked into the vacuum at the beginning.
NU: Which gives us a perfect transition to generative AI. And, you mentioned earlier, authorship issues around generative AI. For our audience who might not be familiar, can you tell us a little bit about some of the big issues there?
MM: Let me start with the caveat of saying there’s a trap that I think all of us have a tendency to fall into, that I try hard to resist. Which is: AI is not a single technology, right?
There's a whole bunch of different technologies that have different affordances, and that are used by people in different ways. I wanna be careful not to give a single answer to this. But we have, for a long time, understood authorship to be a certain kind of activity and engaged in by human beings.
And, one of the questions that is immediately arising? With outputs created by AI, is whether those should count as being authored. Does authorship necessarily entail some level of human judgment and some level of creativity that is inherent in human activity? Or do we just judge it by looking at the output, and saying, “Does it have characteristics that we recognize?”
Like, does it look pretty? Does it sound good? And that's a fundamental question, cause you have to go back to first principles, and ask, “Why do we have copyright law at all?” What are we trying to do with it? And, the standard answer to that, is to give incentives for people to create new things.
People need incentives. AI systems don't. But obviously, the people who create AI systems might want some sort of incentive to create. The Copyright Office has recently released a couple letters about content created using AI tools and the line they appear to be trying to walk. Which in some ways I applaud, because I'm not sure they have the answer right yet.
The aspect of it that I like is that they're starting to get the idea that there isn't just a single answer to the AI question. They're starting to ask questions about what prompts are people giving to the AI tools? And what judgment are they using?
They're more or less saying the AI is a tool - like a camera is a tool. And, we want to know what the humans are doing that are around it - which is sort of implicitly saying the AI is itself not an author, right? The question is, whether there's enough human activity around it to locate the authorship, and the human being.
But then, there's a whole second set of questions. Take some of the image creation tools that are training on big databases of images, and then they're outputting other images. Those are pretty fundamental questions about what it means to copy, first of all? Because some of those tools are ingesting the images, and then training to learn patterns in the images. Frankly when I listen to the technologists talk, I'm not even sure I still understand what they're doing.
They're not doing what you would think of as a photograph. They're doing some other processes on it - and so is that copying? I'm not sure we know the answer to that yet. And even if it is copying, is the training itself fair use?
And then the question is what's the output on the back end? If the system's just learning a lot, should we judge the copyright question as being, is what it spits out on the back end too similar to the front end?
As a way of thinking about that, we could possibly analogize this to Google books, right? Where the Google books project literally scans all these books in. So there's no question that there's a reproduction of the books on the front end - that is prima facie, a copyright infringement. Any reproduction is a violation of copyright.
But then, the Google books project does not make it so you can do full text searching, and then just reproduce the entire book on your screen, right? Or, at least for the books that were still in print, and still under copyright. You would get only these little snippets, and then the ability to go locate it somewhere else.
So, you might call this sort of initial scanning, like intermediate copying. In the service of giving you something smaller. Which the court wound up being persuaded was fair use. And that was very different than intermediate copying, for the purpose of then just giving you the full text on the back end.
I think that is a potentially interesting analogy to thinking about what these AI systems are doing. If you think intermediate copying is potentially okay, depending on what you do on the back end? And that, that's the really important question. That's gonna allow a lot more of these databases to be built, and a lot more of these tools to train on them. If whatever you call that process, that initial training, or ingestion, is itself a copyright problem - that's going to mean there's a big copyright barrier to building these systems in the first place.
—
III. THE PATH TO IP LAW
NU: And Mark, shifting topics, I remember you mentioned earlier that you were an economics undergrad. Can you tell me a little bit about how that's impacted your career? And how you put those skills to use as a teacher, and in the legal field?
MM: A lot of law, a lot of policy–and this is probably especially true around intellectual property–is policy that's created with at least one eye on the question of, “How does this affect various people's economic incentives?”
The ability to think about supply and demand affects the ability to think about incentives, and the way that moving the rules in particular affects different policies. But that's very much “coin of the realm”, in terms of the way people talk about policy. Again, more generally in law, but I think specifically in intellectual property.
And so, I think that an econ background gives you a vocabulary for that. It gives you a way of thinking about effects that people have to pick up - whether that's their training or not. And so, some prior exposure to it, I think, is useful in that respect.
NU: We'll get back to that in a second. Can you tell us a little bit about your career path and what led you to your interest in the field?
MM: I think, like a lot of people say, my career path was in some ways an accident.
I was not a STEM major. I was an economics major but always interested in technology, just not formally trained. I didn't go to law school with any expectation that's what I would work in. I didn't have any lawyers in my sort of immediate, or second order part of my family. So I didn't really have a specific idea exactly what I would do.
I thought pretty safely I wanted to be a lawyer, and I didn't really know a lot more than that. I got a little bit of exposure in law school, but really, I mean, this is what I mean by being an accident. The summer after my first year in law school, I was working at a larger law firm in Chicago and they took us to a social event where I met a person who happened to be working at this little intellectual property boutique in Chicago.
I just struck up a conversation with them and was learning a lot more about what they did at that firm. I thought that sounded really interesting. And I went back to my law school and took as much Intellectual Property as you could. So as much intellectual property as you could at my school meant two classes; IP 1 and IP 2.
If you look at the curriculum that we offer at UCLA, and you think of IP and technology law in broad strokes, we probably offer 50 or 60 courses a year. Different sizes, different dimensions. Some of them are IP specific. Some of them are entertainment related. Some are technology. So, just to think about the growth of that in the last 30 years is pretty astonishing.
I took a couple classes. I actually interviewed with that little firm that this person told me about. I didn't have the confidence to take that job because it was the dot-com era, and all my friends were tripping over job offers at big firms. So I actually went to a different big firm in the summer, but realized halfway through the summer that I had made a mistake.
There's nothing wrong with the other firm that I went to work at. I just realized I didn't want to work at a big firm. I really actually wanted to work at this other firm. And I called them up and basically begged my way back into it and went over to that firm with my tail between my legs. And told them that this time if they made me a job offer, I would actually take it.
And that kind of launched me on this career track because I had a really great practice experience at that firm. I learned a lot from lawyers who were both really good lawyers, but also who had a sort of intellectual bent to them. A lot of them taught as adjunct professors at various law schools around Chicago, but they also were involved in some academic writing.
I didn't know this at the time, but that firm actually was founded by the guy who wrote the federal trademark statute in that field. It was a famous firm. From there about three years into law practice, some of the things I liked best about law school I felt like I was missing. The sort of intellectual exploration; the ability to work on what I was most interested in.
And so I called up some law professor mentors of mine, who basically said, “You didn't do anything right. You're never gonna get a job as a law professor.” And I said, “Okay, but let's assume that I already did those things. And so, now what do I do?” The one thing I did, not for this reason, but that I did right, was that I had kept in touch with a lot of law professors of mine. And so they were able to walk me through how to get in–and I just got a little bit lucky getting a teaching job.
NU: What advice do you have for our listeners who are just getting started teaching courses, and working with students?
MM: This is a really great field to be in. I think we're all lucky to be in it. And, one of the reasons it's great, is because it's ever-evolving. When I think about the academic side of this field - like when I started teaching - the number of people in full-time academic positions at law schools who were working in IP or technology was pretty small.
If I went back–I won't call it a full generation–but if I went back a decade before me, the number was extremely small, right? Like, you could count on two hands who were full-time academics. Now, I go to conferences regularly where there's 200, 250 people there. And those are just the IP conferences. There’s a whole other set of privacy academics and conferences.
And so, one question is, how does the field continually readjust in terms of the way it thinks about itself? And so, when I was in law school, there wasn’t a lot of full-time academic presence in intellectual property. Most thought of it as being related to, or a subcategory of, competition law, or maybe antitrust related.
So a lot of people who taught in IP also taught economics and business law type courses. Obviously that wasn't the real subject matter experts, but that was a way of thinking. And then, the dot-com era happened, and a bunch of other sorts of technology developments. And so, intellectual property became the 800 pound gorilla in the room.
When I first started going to IP conferences, there were a lot of people who were in cyber law or internet law. And those fields were more or less aligned. There were a lot of people doing work in all those areas. Which makes some sense if you think a lot of what was going on was driven by what was happening on the Internet—including changes in IP law, but also more broadly in the ecosystem. IP became such a big field that it actually hived off privacy and data security and maybe even internet law. And they started developing all their own conferences as a cousin field. But now with all the questions about platform regulation - and all the concerns about social media companies - everything's being reorganized under this umbrella, technology law.
The best advice is: be forward looking about what the horizons of the field are, and about what the emerging questions are, rather than getting hunkered down in a narrow sense of what your field is now. Because in just the time that I've been teaching, the tent has changed shape a bunch of different times, and I think it's good for students to understand that.
I have students who come to law school and they say, “I definitely want to be an IP lawyer.” And I say, “Okay, that's great. I can give you lots of advice for that. But I think you should also take privacy, and I think you should take some sort of technology regulation courses. And you should probably take antitrust law.”
You just want more flexibility in the way you think about how these things are all connected to each other, because it's going to ebb and flow over your career.
Listen on Apple Podcasts, Spotify and YouTube.
This AccelPro audio transcript has been edited and organized for clarity. This interview was recorded on April 4, 2023.
AccelPro’s interviews and products accelerate your professional development. Our mission is to improve your day-to-day job performance and make your career goals achievable.
Please send your comments and career questions to questions@joinaccelpro.com. You can also call us at 614-642-2235.
If your colleagues in any sector of the IP law field might be interested, please let them know about AccelPro. As our community grows, it grows more useful for its members.
