Warfare of Art & Law Podcast
Warfare of Art and Law Podcast sparks conversation about the intriguing – and sometimes infuriating – stories that arise in the worlds of art and law with artist and attorney Stephanie Drawdy.
Warfare of Art & Law Podcast
Artists' Rights at Risk: From State Sovereign Immunity to Generative AI
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Cover photo of Rick Allen - copyright Cindy Burnham, Lucky Shot Productions
Show Note:
0:00 Nautilus Productions' Co-Founder Rick Allen gives the history of Allen v. McCrory - suit against NC over Allen's footage of the Queen Anne's Revenge Shipwreck
2:00 NC’s technical arguments against Allen’s claims
2:55 court's use of pendant jurisdiction to wipe out 5 years of Allen’s case
3:45 petition for rehearing en banc denied
6:45 states’ use of sovereign immunity against creators
8:50 Jeff Sedlik’s suit over use of his photo of Miles Davis as a tattoo
9:30 Michael J. Bynum’s suit over Texas A&M University’s unauthorized use of Bynum’s 12th Man book (complaint here; dismissal of copyright infringement claims against A&M employee discussed here)
11:00 Allen’s recommendations to artists to protect their work online
13:35 Emily Gould’s discussion of LAION case
17:00 Bartz v. Anthropic - 23 June 2025 Order on Fair Use in N.D. Cal.
19:50 Allen on opt out policy
20:20 Visual Artists Copyright Reform Act (VACRA)
21:00 Gould on survey by DACS (the Design and Artist’s Copyright Society)
23:00 response to UK government’s consultations
24:00 UK House of Lord’s hearings
24:30 Allen on artists not understanding impact of generative AI
26:00 Gould on UK judgment from trial in Getty v. Stability
28:50 Gould on judgment in GEMA v. Open AI
31:55 Lauren Stein on ChatGPT and law school's encouragement to use AI
33:00 Getty v. Stability in UK – Getty’s drop of direct infringement claim and appeal of ruling on secondary infringement claim
35:50 UK’s Section 9(3) - copyright protection for original work created by a machine
39:20 Stein on copyrightability and Japan’s approach to sufficient human authorship
41:40 Gould on Beijing Internet Court's judgment in Li v. Liu
42:05 Allen’s position on AI
44:00 Gould on authenticity and human contribution
47:35 Ed Newton-Rex
Please share your comments and/or questions at stephanie@warfareofartandlaw.com
Music by Toulme.
To hear more episodes, please visit Warfare of Art and Law podcast's website.
To leave questions or comments about this or other episodes of the podcast and/or for information about joining the 2ND Saturday discussion on art, culture and justice, please message me at stephanie@warfareofartandlaw.com.
Thanks so much for listening!
This podcast and its content may not be used for training or developing AI systems without permission.
© Stephanie Drawdy [2026]
Blackbeard Footage And State Takeover
SPEAKER_00I have a collection of footage from Blackbeard the Pirate Shipwreck that I own, have copyrighted and registered. The state of North Carolina decided to make it theirs, passed a law to convert my footage to the public domain. We filed a lawsuit in 2015. We have been up to the Supreme Court once, back to the Fourth Circuit multiple times, and may still yet be headed back to the Supreme Court again to try to protect our intellectual property rights.
SPEAKER_04Welcome to Warfare of Art and Law. The podcast that focuses on how justice does or doesn't play out when art and law overlap. Hi everyone, it's Stephanie, and that is Nautilus Productions co-founder Rick Allen, describing the history of his lawsuit against the state of North Carolina over its unauthorized use of his underwater footage of the Queen Anne's Revenge shipwreck. Rick was previously a guest on the podcast at episode 99 and returns to a recent second Saturday to discuss with others not just the impact of state sovereign immunity on artist rights, but also the impact of emerging technologies like generative AI.
CRCA And The 9-0 Loss
SPEAKER_00In 2021, the Superior Court, after we lost at the Supreme Court, we used um the CRC Ray or CRCA or Copyright Remedy Clarification Act as our justification for suing the state of North Carolina. They claimed that they had sovereign immunity, as you would uh well guess. And um so under the CRCA, we petitioned the Supreme Court and we were unanimously shot down by the Supremes, lost a 9-0 decision because they decided that the CRCA was um unconstitutional. So we went back to the district court in North Carolina in 2020. 2020. By 2021, the district court judged uh on some very, fairly narrow issues, one being a Georgia argument um in takings, uh said that we could uh refile, well not refile, but uh adjust our lawsuit. Um and so we did that, we amended our complaint, and the state immediately appealed on multiple uh points, and we ended up in the Fourth Circuit at least twice on based on their technical appeals of the ruling against sovereign immunity, and also whether we should have used rule 54 or rule 60 to amend the complaint. And um, so we were just heard this pack past October in the Fourth Circuit over that issue. And again, the Rule 54, rule 60 issue, and sovereign immunity once again with the state. And the Fourth Circuit uh in a 3-0 opinion by the panel shot us down again. And instead of considering the two issues on appeal, um went against their own um uh precedent and the Supreme Court's own precedent and said and used what's called pendant jurisdiction and basically wiped out our last five years of litigation. And so last Friday we filed um an appeal and asked for an en banque um review of the ruling, and so that's where we stand.
SPEAKER_04And what I was reading about that is that the Solicitor General was
Fourth Circuit Maneuvers And En Banc
SPEAKER_04making the argument that you can't have this do-over, and that that even that Georgia case that you mentioned, it's existed since 2006, I think. And so why are you allowed to bring it up now? So, do you want to address that perhaps? Sure.
SPEAKER_00Um, I think the first thing is that when you go to the Supreme Court, as you well know, you've got to select an issue on which to stand. And so at that point, we based our uh case on the fact that the Copyright Remedy Clarification Act was a standing statute and that we would use that as our tool to uh get a ruling. And you don't consider at that moment that a statute passed by Congress and that had been a law for at that point uh almost 30 years would be ruled unconstitutional. So, how could you even foresee an argument over uh a Georgia complaint? And so, and also during the Supreme Court argument, the Solicitor General for the state of North Carolina conceded that we had a valid argument under the Georgia theory. And so that's one of the reasons that the district court judge allowed us to refile. So they opened the door for that.
SPEAKER_04So what are your thoughts with this cert petition?
SPEAKER_00Um, well, first off, we've got to wait to see if the on banque is uh where that goes.
SPEAKER_03Yeah, well, I feel like not to be negative, but I'm like, I feel like maybe I know where that's going. I hope not.
SPEAKER_00Well, to say that me and Sisyphus have a lot in common would be an understatement. Uh I would be really surprised if they do uh either take it, take it to look at it or reverse the lower panel. Um but as I said, the three-judge panel uh violated their own precedence and Supreme Court precedence over this issue. And so that gives us a pretty good reason. That gives us a circuit split and a uh Supreme Court uh is a Supreme Court split as well. So we we might have a good argument. And the really, I guess the frustrating part for me is we are still fighting a technical battle after 11 years. We have yet to actually address the facts of the case or get an answer from the state of North Carolina over from our amended complaint. And so they, you know, they have spent every bit of their energy on trying to find a technical or legal reason to throw this case out and not address the actual issue, which is them passing a law to take my
Why Sovereign Immunity Threatens Artists
SPEAKER_00property.
SPEAKER_04So that is I I think the bigger message that I always am stunned by is one, that they are being so vocal about how they're just steamrolling over your rights, and two, that there aren't that many artists that know it it seems it's it's really complex.
SPEAKER_00And I think the big issue is really sovereign immunity, and if we chip away at that, it it puts a crack in a big door in a lot of sovereign immunity cases, you know, when officers shoot somebody and the state claims oh they've got sovereign immunity. Um, you know, look at Minneapolis right now. I don't think any court really wants to deal with that. But on the flip side, how and and as we've discussed this many times, you know, how do you have a a copyright scheme where states can just take your intellectual property without any uh license or agreement and do what they will with it? And if I turn around and say took the state's copyrighted material, I would end up in court at best.
SPEAKER_04Yeah. So as artist rights issues are coming to a head from many angles, this is one that's just been under the radar for so long. And I appreciate so much the energy that you and your legal representatives have put into this case because it is so important for so many other people.
SPEAKER_00Yeah, it is, and I have a fantastic legal team, you know, they they are all in and have have been from the beginning. And then, like you say, you know, look at this in the uh in the pretext of uh AI and what that's gonna mean for all of us as artists. That's complicated enough. But then if you have state entities using AI and slurping up your intellectual property, what are you gonna do? So somebody somewhere has got to deal with this sooner rather than later.
Other Copyright Fights With States
SPEAKER_04There were, when we last spoke, some other cases that had been pending in the US that you had connections with some of the plaintiffs. And I just I wondered if there was an update on any of those that you might want to share, or how any other cases that have been kind of satellite cases around yours have been progressing.
SPEAKER_00Yeah. Um, the two that come to mind are Jeff Sedlick, whose um photo and I can't remember the jazz artist, but he took a photo of a jazz artist and then a tattoo artist put that tattoo on somebody. And that's been Miles Davis, I think.
SPEAKER_01Miles Davis. Exactly.
SPEAKER_00Yeah, Miles. Thank you, Emily. And so that's been working its way back and forth through the courts. Um, I think he just lost another appeal. Um, so I you know that's another complicated issue because I've certainly seen the photo and the tattoo, and I can't tell a difference. So um that's going on, and then Mike Bynum, who was suing Texas AM University for uh misusing um his book about the 12th man, and that is still going on. Um the judge did rule that the person who was in charge of public information for Texas AM at the time was liable and not immune. Um, that person has since appealed that again, even though the Fifth Circuit said he was he could be sued. So he's still chugging along too with his lawsuit. But um just you know, whenever you're dealing with state entities, it's it is challenging at the very best. So those are the two I can think of.
Practical Steps To Protect Work
SPEAKER_04What are the suggestions you have, knowing we have these kinds of concerns about state sovereign immunity and concerns, as you mentioned, about emerging technology and the issues that are coming to the fore for artists with that? Did you have any ways that you have been protecting your work? Even websites, how to protect your sites. Like there was a recent decision that was clarifying that when you have robots TXT on your site, that's just an instruction, a request. It's not a command, so it's not a real lock on your website. So it doesn't trigger a DMCA violation.
SPEAKER_00Yeah, well, no, and you're absolutely right. I mean, I've got I think 24 different anti-robot text embedded in my website, you know, for what good that does. Um, first off, even though we lost and this copyright remedy clarification act was declared unconstitutional, that didn't invalidate copyright law. So um, you know, the act of copyright still exists, it still grants you rights. Um, you still register your work, and in theory, at least uh with everybody else but government agents, um, that has some weight. Uh, watermarks are still um useful. Um, you know, big picture, you know, if you're dealing with a state, uh my answer would be don't. Because, you know, in every state contract, buried deep within it is a sovereign immunity clause. Um, my wife's a professional photographer and was doing some work or wanted to do some work for a local university, a state university. And we went back and forth over the contract because as you go deeper and deeper into it, there's a paragraph buried in like page four about in the end, state has the state has sovereign immunity, and regardless of what anything else in this contract says, you know, we can invalidate it. And I'm sure that is not just North Carolina that does that. So as an artist, I would be very wary of any work with the state right now. It's not what I can't there's no legal protection that I'm aware of. And I know that soon after we lost in the Supreme Court, I believe it was a software developer in Indiana brought a case against the state of Indiana, and the judge in it said, Have you not read Alan V. Cooper? And threw the case out, and they were done. So um, as to AI, uh, same deal. I know how do you stop these, you know, multi-billion dollar companies from scraping your work? And you know, fortunately that's working its way through the courts. Um, I was listening to NPR today, and there was uh something uh related to that. But um I don't know, it's a tough time to be an artist and trying to protect your work. And you know, we have we live in a world where we want our art to be out there, that's the entire point, but we also deserve to make a living from it, just like everybody else does. And I don't uh you know, I there's a huge dam here that's going to break, and I don't know how it who or which way it's going to break.
Europe’s AI Scraping Opt-Out Debate
SPEAKER_01Stephanie, you were just saying, I was just uh thinking, you were talking about the the using the robots TXT to try and protect your work when it's online. There's been an interesting development in a case in Europe quite recently. So the Lion case, I don't know if anybody's been following that. So Lion, this company that uh well it started out as a well, I think it is a not a non-for-profit, but it's quite closely connected with other companies like Stability AI, which is a commercial company. But so they created this database of effectively URLs, which then pointed to images um to use in AI training. And they sort of yeah, they got a favorable judgment. Um so there's so in the US, most of the discussions are about fair use, but we don't have the same system in um in the UK and Europe. And in Europe, there are something called a text and data mining exception, which allows AI companies in some circumstances, and this is all being debated at the moment, but in some circumstances, to use uh to basically just scrape data and use it under these exceptions, which were originally for um to allow people to do sort of scientific research without having to get copyright licenses. So, anyway, um there's been a lot of debate about whether they apply or not, and and there was this case against Lyon, which basically said the initial verdict was that um the the kind of collation of those URLs and the creation of a database as a sort of non-commercial exercise would fit within these exceptions. So they could do that. Then there was an appeal, and the appeal failed on the main issue, but there was quite a bit of discussion about um how, as an artist, you can opt your work out of this scheme. So you can say, okay, well, there's this exception, but I don't want my work to fall within this exception, I want to protect it. And um in the initial judgment, as I understand this, they um there was a sort of discussion about the fact that, well, actually, the way in which you can exercise that opt-out might be quite broad, might be quite wide, and you might be able to do that by simply having something in natural language, you know, in words on your website to say you can't take my you can't take my word without consent. But in this appeal decision, again, this is my understanding, I haven't read all the detail of it, but my understanding is that they have now said actually it's gonna have to you're gonna have to um uh opt out by uh some form of uh machine readable, sort of you know, automated form of opting out. So it's making it more difficult for artists to, you know, it's more technical, they've got to go through more hoops to protect their work in that way. That was one thing
Fair Use, Piracy, And Registration
SPEAKER_01I was thinking of. Then the other thing, um, in relation to some of these AI cases, um, what you were saying, Rick, about registering your copyrights. I mean, again, we don't have that system in the UK and Europe, but it's really important, and I think it's gonna be increasingly so, for US artists and US rights holders to register their copyrights in the context, and I was just reading about this earlier today in a totally different kind of context, but um, of one of those US cases involving Anthropic, the company Anthropic, where um they so very, very much sort of summarizing what I understand of that judgment, anthropic. There was a fair use defence whereby um the court held that sort of broadly speaking, the use by AI companies of copyrighted words for their training is transformative. And I think that was the case where they sort of analogized the way in which an AI tool is developed with sort of the human brain, you know, it learns in that way, it's just learning like a human brain, and you know, nobody has stopped you know from using their brain to kind of um you know use someone else's idea, or um, so uh what was I saying? Yeah, so so that was the basic premise, but then um the court also said that that doesn't work if these AI companies are using pirated copies, and in fact they had done, they'd used quite a lot of works in these kind of what do they call them shadow libraries, and so although Anthropic had a sort of partial victory in that their fair use defence worked to a certain extent, because they used these pirated copies just before there was going to be, I think, another hearing about damages, they settled. And it was a massive sum. The global sum was massive. I can't remember what it was, but I remember reading that it was the biggest uh settlement of any US copyright case ever. But finally getting to the point here, um, that meant that the individual copyright owners um of all of those works that they've used that have been pirated, they have a claim for, and it's it's it's like a uh an overall sum, and it's not massive, but it's something like $3,000 or something. But only that is only gonna work for them if they have registered their copyrights. So there might be loads and loads of people out there who would just not think to do that, you know, they're not kind of the whole legal side of what they do, and then they will lose out. So all that is to say that registering your copyrights, I think, is going to become increasingly in important in this in this age of AI and scraping. Sorry, that was all very long, but it just struck me those points are kind of you know, all of this is part of one huge big discussion at the moment, I think, isn't it?
SPEAKER_00I'd say you nailed that, Emily. I I just I love the whole opt-out concept. It's like saying the only way to stop you from burglarizing my house is if I opt out of criminals doing that. It's ridiculous. It's problematic. Yeah, I agree.
SPEAKER_04Yeah. Yeah, and I definitely echo both of you your point about copyright registration. I've been thinking a lot about this because there are so many people, myself included, who have in the past not really taken that as seriously as
VACRA And Policy Moves Abroad
SPEAKER_04needed. And there were many reasons why, one being the price of an application. And that was one thing that I was excited to see that at least now there's been a bill introduced that would I this is my favorite part of this bill. It's called uh VACRA, so it's the Visual Artist Copyright Reform Act of 2025. But one of the provisions would give you a subscription so you could just pay one fee and then register however much you wanted. I love that idea. That would be a way that would really make it affordable.
SPEAKER_01That's brilliant. How far has that bill got, Stephanie? Do you know?
SPEAKER_04Is it not very far. I think it was like the end of last year, and so it's you know, struggling along. But hopefully it's a way to also argue to the artist community, you really need to start protecting your work formally. In the EU and the UK, are there more proactive steps that you see artists needing to make, Emily?
SPEAKER_01Yeah, that's a really interesting question, Stephanie. Um so yeah, it's strange, isn't it? So in the long distant past, there was a copyright registration, I mean centuries ago. Um, but because of you know the Bern Convention and the idea that it's an informal right, you know, that was that was all done away with quite some time ago. Um but I don't know, it's it's a good question. I don't honestly know the answer. I know that's the collecting societies are doing a lot of work, a lot of lobbying work, DAX in particular. They did a really big survey a couple of years ago, and it it was interesting. I was hearing artists talk about it recently. Um because it is so difficult, isn't it? I mean, artists are very enthusiastic about the potential creative opportunities. Of you know using AI in different situations, um, but also you know very fearful of what it means for uh their livelihoods and you know in other situations. And the I think that the headline points from from the DAX survey about what artists are really worried about is the transparency issue, the fact that they just do not know. I think some in the 90s of percent uh of artists who, even though there are, you know, there are these tools, have I been trained and those kinds of things, where you can look to see whether uh your work has been scraped, uh, lots of artists either don't know how to use them or haven't thought of using them and just just didn't didn't know, you know, felt like completely powerless and in the dark as to how they might go about protecting their work and how they might know whether whether or not um it's been scraped. And um so yeah, there are certainly lots of concerns, but in terms of sort of positive steps, I think it's at the moment I think the efforts are mainly directed towards the lobbying and the trying to um trying to convince the government of the need for protection and a balanced approach because the government in the UK has has really sort of flip-flopped um initially the with the the Conservative government sort of two or three years ago when when the discussion started was very, very pro-business and was trying to expand all of the exceptions to enable AI developers to um you know have kind of carte blanche, really. And then there was a massive backlash from the creative industries, um, which you know are economically worth a lot of money in this country. And um, so there was one consultation and then another consultation. Um, and so at the moment we're sort of waiting for legislation. So there's a lot of political activity going on. There are several different groupings within, particularly within the House of Lords, actually, is really interesting committee that's been hearing a lot of evidence from academics, from artists, from publishers, from um uh musicians, and um and they they're really you know getting to the heart of the arguments, and so um, yeah, we're sort of waiting to see what comes out from from the results of this consultation. And I think there will be legislation in this country within a couple of years, maximum, I would say.
SPEAKER_00Yeah, I was I was just gonna say, I mean, you both bring up a great point, is that artists just really don't understand this and the scope of it and what can be done.
The Hidden Cost Of Using AI
SPEAKER_00And now virtually everybody's an artist. You know, if you've got a feed on TikTok or whatever, you're an artist, you know, whatever you're doing. And the other thing that that really concerns me, and I am not an expert on AI, but even these AI artists who use their own work and then manipulate them to create other transformative things, you're putting your own artwork into an AI program to then make something more with it. But you've provided your artwork to that giant database, to that giant animal now, and it becomes part of everybody else's stuff. And I I don't know, you can't protect that either. So just in using the tool, you are giving up your creativity and your tools and your property to make something else. And for those of us who make a living from our art, I just that's not a good trait. And I don't, I don't think, I don't think anybody really understands that. You know, Adobe says, oh, you can work within Adobe and you know it's all safe, but Adobe is hoovering up your artwork if you put it in to make something from it. And uh I I really find that disturbing.
Getty, Stability, And Infringement Puzzles
SPEAKER_01Yeah, it uh it's totally fascinating how these issues are being dealt with by the courts in very different ways at the moment, because that you've probably heard about the Getty case, a Getty instability case. It's been it's the first um AI copyright case that we've had in the UK, and the reasoning, I mean, the judgment there, I I I read, I waded through it, it's huge, it's about 2010 pages. Um, and I have to confess I did not understand all of the technical evidence. My goodness, it's very clever stuff. But the the the sort of headline from that verdict was that um there essentially wasn't an infringement that so the stability tool wasn't an infringing article because it does the way in which it works means that it doesn't actually copy or store the underlying images that it uses to train. So it's very difficult from a legal perspective, from a copyright perspective, to identify what the actual infringing act is, what the infringement is, because what the tool does, it's fed all of this data and it um it learns from the data. They kept using this word in the judgment exposure, so it's exposed to that data, but they were very careful to avoid saying, but you know, it it copies or reproduces or stores the data, which would you know be infringing acts. So you're left with this kind of conundrum, I think, whereby as the law currently stands, it's very, very difficult to apply it to that scenario and identify a clear copyright infringement. But sort of ethically, I suppose we might say, it doesn't feel right that that tool, that model, that AI model, only works in the way that it does because it has in some way used that data, and that just doesn't feel fair. It's used that those images, those those copyright works that you know people have expended their creative energy to create, it's used them without consent, without compensation, without remuneration. Um, and there's very little from a legal point of view that the copyright owners can do singingly about it. So that was one judgment, that was one way of looking at
GEMA, Memorization, And Direct Copying
SPEAKER_01it. Then almost around the same time, there was a judgment in Germany uh which came to a very different conclusion. It was so, and this is one thing I think is really sort of interesting, and it will be interesting to see how it moves forward. Different kinds of works, I think, throw up different sorts of problems. So images are one thing, then we've got music and text, and the the tools, the algorithms for those different when when they become generative tools, they operate in different ways. So the the GEMA case uh in Germany, so GEMA is the collecting society for the music industry in Germany. So ours is called PRS. I'm sure I don't know what the uh the US one is, but so so it's the it's the society where you go to if you are a musician to collect your royalties for whenever your music is is played or performed. And that this is so I I think what happened in Germany, I don't know the case massively well, but I think that society tried to get a licensing deal with Open AI for use of you know its its um uh its musicians or the the you know its rights holders uh compensation so it could you know get a deal with them so that it could then pass on royalties for to all its musicians. And it wasn't successful. AI OpenAI for I don't I don't know how far the negotiations got or what they said, but anyway, they couldn't get this licensing licensing deal. So they basically brought this case as a test case, really. And um they uh so what they did was identified nine songs where the lyrics were directly reproduced by Chat GPT, and so they you know they asked, they put in a prompt, you know, what are the words of the song or blah blah blah, and it was able to immediately reproduce the words, and so in that case, the court ended up saying, Well, look, what has come out of this tool is the an exact reproduction of the words of those songs, and so that tool has somehow got to have memorized, and they keep using this word memorization, it's it's kind of become a bit of the buzzword of how these cases are decided, but so that tool has somehow memorized those words, and um even if you know we can't sort of see how they are sort of in any tangible way stored within that, if that's what the model has done, then that is a reproduction and that is an infringement. So a very different way of coming at that same kind of debate from the UK court, which took a very sort of technical, technical way of assessing it. But yeah, so yeah, different results, and so who knows, you know, who knows what the next case is gonna say.
Jurisdiction And Secondary Infringement
SPEAKER_04That's the biggest thing I think is really compelling is that this difference of understanding how the technology works and who's getting it right. Exactly.
SPEAKER_02I was reading a thing that said on chat GBT, you can toggle off in your settings if you want your entries to be used to train the program. And I don't know how much I actually believe that that is not true. And I am in law school right now and we use AI, and our professors encourage us to use it for drafting or in different ways to kind of, especially um Lexus and Westlaw AI, definitely encouraged to use. Um, so it's interesting that these companies are saying you can select to not have your data for training, but that's very, very iffy. Not sure about that.
SPEAKER_01That's interesting that you're encouraged to use AI. Yeah, really interesting. Because, you know, it's I I was hearing someone uh talk the other day about the fact that um school kids in school are gonna be taught sort of prompting. What's the best way to prompt an AI school? That's gonna be, you know, it's gonna be class, so that's that's that's what it's gonna be.
SPEAKER_04Going back to the idea in the UK decision that training wasn't happening in the UK, and that was a really big deal, but should that be? We've said this a lot, Emily. Like, are the current laws fit for purpose? Are the current interpretations of the statutes fit for use in applying to generative AI contexts? So, like in the UK Getty decision, the issue uh that's going up on appeal about secondary infringement, yeah. Tell me what your thoughts are on that.
SPEAKER_01Yeah, yeah, no, absolutely. Yeah, so so you're absolutely your point about jurisdiction is absolutely right. So that so Getty had to drop the um primary infringement claim at the very last minute, actually, but um because it became clear that almost none of the development or the training was happening in the UK. But you're you're so right, Stephanie. I mean, it just doesn't really make sense, does it? It makes everything so arbitrary if you can just you know move people a couple of miles outside the jurisdiction and then you're fine. It just really doesn't. And the actually the EU AI Act tries to get over that, it tries to have this extraterritorial reach. Um, and then what oh yeah, the about the appeal. So, yes, it's a secondary they're they're appealing the the secondary infringement claims. So the a lot of the discussion uh in the decision was about this definition of what an infringing article is, and the statute um is from 1988, so you know, well before we were thinking about any of these kinds of issues, um, and and yeah, it was quite startling to me actually that so so there was discussion firstly about what is an article and can software, can an algorithm, a model, be an article, and it was apparently a groundbreaking point that um that the judge accepted that that that could be the case because previous cases had always required some sort of tangible object, like you know, uh a USB stick or some sort of hardware, which I mean I was quite stunned actually that that I was surprised that that was the case, that that discussion hadn't been decided before. So that was the first thing software could be an article, and then um the discussion about whether whether it was infringing, but that was you know it's all one kind of composite phrase. So yeah, you're right. There was the they you know, they in the case they looked at um sort of the history of how those phrases have been interpreted, and you know, what was said at the time when the statute was made. And so, yeah, so all of that will will be rehashed in the appeal, I think.
Machine-Made Works And Human Authorship
SPEAKER_04Well, I I guess uh too, I'm wondering uh just a survey, and we'll start with you, maybe Emily. In the UK, this question uh with all of the consultations that are going on, I know one of the issues was about Section 93, copyright protection for machine-generated works, and whether or not it should stay on the books, it hadn't been used very much, and whether or not it should even apply to generative AI. So does that have any movement or is it still kind of because I've seen a couple of criticisms that were formally put out that the consultation had had included.
SPEAKER_01I my hunch is that it will go, actually. Um, but I I'm not a massive expert on it, so that's uh that's probably a fairly uninformed hunch. Um, but what I've I mean, I don't think it's been used very much. I don't think, you know, I don't think it's been known about very much. Um and so I I think um there is a sense that I think it's always been a bit problematic. So this so there's there's this right in the UK that says that um you can have copyright or a limited copyright in a computer generated work. Um, but it's a very the way it's drafted, it's very confusing because to get copyright in the UK, your work has got to be original, and that means it's got to be the intellectual creation of an identifiable author. And the point with a computer-generated work was that it's a work that is created sort of in the in the absence of an author, and yet it's supposed to be original, so there'd always been this kind of internal conundrum. Um, and nobody quite, I think, understood what it, you know, all the benefits of that um particular provision were. So um, yeah, I think, and in the context of AI, even if you had that right, that right belongs to um the people who uh sort of I can't remember the exact phrase now, make it possible for the words to be created kind of thing. Um, and how do you identify those people? Um, when there is output from uh an algorithm, who are the makers, who are the creators, who are the people that should benefit from that right, from that exploitation rights? Is it just the people who prompt that uh algorithm to create, I don't know, a dog in the style of Picasso, or is it the people who have developed the coders who develop the code, or is it the people who've collected together all the training data, or is it the people who have uh adjusted the tool when it learns from the training data? It's such a it's a minefield, I think. Um so for a number of reasons, I yeah, I think it will probably go.
SPEAKER_04And so in the US, we have very clear bedrock human authorship standard for copyright. And I was just curious thoughts on that, Rick and Lauren, both of you being creators.
SPEAKER_02I would. I oh no, Lauren, go ahead, please. Oh, thanks. I actually wrote on this for my journal last semester about what constitutes sufficient human input because it's not defined by the US Copyright Office. So I took a look at international standards and I actually really liked the Japanese approach. So they have four specific prongs that they use to analyze. So kind of drawing on that and applying it to the US, I did a two-prong test and it fell in line with China, the UK, EU, basically all other countries, but provided a little more guidance, which I think is beneficial for the courts. But I was really intrigued. I don't know if you guys followed the cadre case last year. It was in California for Emily. It was not in the US, it was versus Meta. And the court, the Northern District of California did hold that it was sufficiently transformative to train LLMs, which I thought was really interesting. And the plaintiffs couldn't show that it impacted their market. So I think that's really interesting that they didn't say using it was wrong. It was a showing of market infringement was required. So I think that's really interesting. And I don't agree with that. I think that using it for any purpose is wrong, even if you can't necessarily show market infringement. I just thought that was a very interesting case. And Rick's nodding his head, I'm sure you are well aware of that case as well.
SPEAKER_01Lauren, what's your kind of what's your two-pronged test that you proposed for the creativity? Because Japan um is quite a sort of pro-developer regime, I think, isn't it? So, but they've clearly sort of taken into account creatives' concerns as well.
SPEAKER_02Yes. So I did my first prong was human additions and corrections. So I thought whether the human author made creative additions and modifications or corrections to the AI-generated output such that the human contributions outweigh the AI system's autonomous work. And then the second one was a mountain content of instructions and input prompts. So that was a level of detail, creativity, and control exercised by the human author. So right now they weren't exactly clarified by the US Copyright Office. They sort of hinted at them, but they never propose a specific test for sufficient human input, which I think is critical for the courts right now.
SPEAKER_01There's the Beijing case, isn't there? That was quite a while ago where that there was the the was that was a social media reposting, wasn't it? I think. Um, but the the claimant there, the rights holder there was successful on on that kind of basis, right? At mind something. Rick, did you have any thoughts you wanted to share?
Rick Rejects AI And Market Value
SPEAKER_00No, I'm just glad Laura chived in. That's really good. Um, you know, I I avoid AI, and I I probably need to understand to use it, but I just I'm fundamentally opposed to how it works and what it does. And I do have the luxury of not having to use it. So it's it's it's pretty easy to take that stand. But you know, when it comes to market effect, like I have a very narrow set of things that I license that nobody else in the world has. And if I ever use AI in conjunction with those that footage or whatever, it's gonna be lost to the world, and I will wipe out any value it has. Um so that's that's very concerning. And you know, if if you if you're an artist who has a very unique style, I just even getting close to AI would seem very dangerous to me. Um the other thing that I I will predict in the future, and it may not happen while I'm on this earth, but I think there will come a time when art will be certified non-AI, and there will be a value in it because an actual potter or musician or uh painter creates something, or even a digital artist, and they can certify that no AI was used in it, and it will be their very own creation.
Certified Non-AI Art And Trust
SPEAKER_00And I think there may come a time when there's value in it. Right now, you know, these multi-billion dollar companies are have found a way to basically engage in the theft of data from everybody on the planet and monetize it to their benefit. And you know, you see these things where, well, we're gonna license things from artists, but I've yet to see anything that puts a real dollar figure on what that means. Like if you're if your images are used to build this other thing, is it gonna be like Taylor Swift who gets a few pennies for the playing over a song? And is that fair? Um, you know, if you look at artists and musicians, they don't get compensated fairly for their work. I can't see the rest of us being compensated either. So I just uh it's just it's just theft.
SPEAKER_01It's so interesting what you say, Rick. I'm doing some research at the moment about um sort of authenticity generally and what and what people feel is an authentic word versus something that's not genuine. And I'm thinking about it a little bit in the context of AI. And I think you're so right that um there is going to be value placed on and associated with something that is that clearly has human input that is clearly a human expression as a kind of an antidote, as a reaction to the fact that we just don't know what is real in this day and age. We have no, you know, deep fakes, and it it's it's uh I think it's almost getting to the point. It's funny. My my husband showed me a photo on his phone the other day, it was about some press story, and my immediate reaction was, Oh, come on, Paul, that's not for real, that's clearly AI generated. And it turned out it was actually a real picture. But it I think there it there is, you know, that you do increasingly, I think, question everything you see, and so that idea that the this you know that you can can have some sort of certification of something that is real, that has come from a you know, from genuine human creativity, is uh will be a thing. And in fact, I was reading something, I won't talk much about it because I might get it wrong, but it was a while since I read it, but it was about about on this point whether so there's something called there's a legal protection. Um, it's a little bit like a trademark, but it's called a GI, uh and an indication of geographical origin, and you have it for things like um, I don't know, like artisanal produced things like for champagne or something that's got to have come from a particular place. And the author I was reading, I'm sure this is the point. They were saying, well, could we have something like that to certify exactly what you were saying, to certify something which is of human origin, which I thought was a really interesting idea, it might be hard to work out in the practice, but um, this idea that you know that people are gonna value things which um have genuinely come from an individual, a person, and so if there's a way of sort of legally certifying or protecting that, that might be something that's quite interesting.
Ethical Data And A Cleaner Future
SPEAKER_04So we start off today with talking about RIC and state sovereign immunity and go over to generative AI and and the whole arc is rather bleak on some level.
SPEAKER_03Any closing thoughts by anyone?
SPEAKER_01Just very quickly, Stephanie, on that point, um, slightly less bleak, um, that point about that point about um you know uh sort of trusting the data and and and uh you know the the theft of data and and and and how this might be controlled. There are, I think, um there is a move towards um uh people wanting to know where that data has come from and if they're using a tool that that tool has been created and developed ethically. So there's a guy in the UK who he started out with Stability AI and he's almost kind of gone turncoat in that he has now set up a company that is all about creating databases and ethically. Um, and so I think increasingly companies um are commercial companies are are not are not gonna want to take the risk of you know lawsuits from rights holders so that they are going to there is gonna be a market for clean data, I think. So perhaps that's a little bit less bleak.
Links, Reviews, And Farewell
SPEAKER_04There will be links in the show notes to learn more. If you were intrigued by this podcast, it would be much appreciated if you could leave a rating, a review, and tag Warfare of Art and Law Podcast. Until next time, this is Stephanie Droddi bringing you Warfare of Art and Law. Thank you so much for listening. And remember, injustice anywhere is a threat to justice everywhere.