Shauryaveer Chaudhry & Aastha Nayak

Hosted by Aastha Nayak and Shauryaveer Chaudhry, Senior Editors at the Law School Policy Review, this podcast examines India’s intellectual property framework. The discussion draws on Prashant Reddy T.’s Create, Copy, Disrupt: India’s Intellectual Property Dilemmas. It explores the relationship between IP protection and innovation. It then turns to generative AI and copyright. The speakers discuss text and data mining, reproduction, compensation, and copyright exceptions. They examine the idea-expression dichotomy and authorship of AI-generated works. The conversation also considers traditional knowledge and geographical indications. It concludes with Basmati and emerging questions at the intersection of agriculture and intellectual property.

LISTEN TO THE PODCAST

Shauryaveer Chaudhry: Hi all, welcome to the LSPR symposium on OpenAI versus ANI media where we examine the broader legal and policy questions raised by the ongoing proceedings before the Delhi High Court. The symposium considers what India’s copyright and information ecosystem should look like in the age of generative AI without inviting submission of the merits or the outcome of the pending litigation. I would now like Aastha, our senior editor, to introduce our guest for us.

Aastha Nayak: Yes, our guest for today is Mr. Prashant Reddy T. He is an alumnus of the National Law School where he graduated with a BALLB honours in 2008 and he later got his masters which is his LLM in Law Science and Technology from Stanford. He has done a lot of research on IP rights. But he has worked with law firms as well, so we believe he has a unique insight into how the law operates in both theory and in practice. Today we will be discussing his book ‘Create, Copy, Destruct: India’s Intellectual Property Dilemmas’. I would also highly recommend his newer book, which is ‘Tariq Pi Justice’ on reforms for India’s district courts. I think anybody who’s interested in public law would like his new book as well, but for today, we will focus on IP. So I will hand it over to Shauryaveer to introduce the topic of the project podcast.

Shauryaveer Chaudhry: Right, so the topic of the podcast is obviously sir’s book ‘Create Copy Disrupt: India’s Intellectual Property Dilemmas’ where independent India has inherited its intellectual property law from the British Raj, which was in time obviously shaped according to the realities of our country by the courts and as well as the legislation where we navigate international treaties in line with the changing national interest.

These changes affected everyday life. They shaped medicines, music, books, movies, food, yoga, and the Internet. This book kind of tells that story. It traces the development of Indian IP law and policy in modern India. It focuses on different industries and sectors. This book draws from many sources. These include parliamentary debates, export reports, interviews, archival material, and case laws. It focuses not only on the black letter of the IP law and exposes the politics and history behind Indian IP law and policy. Sir, thank you so much for joining us. And now I would like Aastha to ask the first question.

Prashant: Thank you for having me. Pleasure to be here.

Aastha Nayak: Thank you. So you wrote your book, it was published in 2017, so I am assuming it’s been about a decade since you wrote it. Do you think India has finally favored the Imitation Policy in IP or does it still stick to the old Innovation Model? In your book you describe in the introduction, the title, which is ‘Create, Copy, Disrupt’, is a deliberate departure from the conventional phrase ‘create, protect, innovate’ in IP policy. You write that a developing country like ours might actually benefit from a policy of imitation, even if that imitation is disruptive. So looking back almost a decade later do you think that we have embraced this idea in practice, or has Indian IP policy moved back towards a more conventional protect and innovate model?

Prashant: See the general picture of the last 10-20 years, ever since the liberalization process started and India is moving towards a free market economy, partially due to the WTO and trips, etcetera…There has been a lot of pressure on India to comply with developed countries’ standards of intellectual property, and there is, you can see that in rhetoric coming out from the Sarkari economists and senior bureaucrats where they will talk to you about the number of patents granted in India as a sign of great innovation, and everybody who is in patent knows that the number of patents being granted is absolutely it is not a sign of anything, right? Least of all, innovation in your country, because a lot of these patents are being filed by entities outside India. I mean everyone just has to secure a patent in India and it is quite easy to do. So those numbers are always going up. Sometimes they try to show even the domestic patenting number that is filed by Indian entities and companies, even those numbers are going up. But generally the number of patents granted is not a very good sign of whether there is innovation because the brand office basically will give a patent for almost anything unless it is, like, absolutely ridiculous. Even like in the US, there’s a famous case where they gave a patent for a peanut butter and jelly sandwich at one point. And now another patent for laser pointing device to exercise a cat, basically you wave across the laser device and the cat will run after it. So band offices do all those kinds of crazy things and policymakers are enamoured by these kind of numbers and they buy a lot into the rhetoric of strong IP will automatically mean that you have a more, like, a more productive economy or a more innovative economy, which is one of those stereotypes that I think our policy makers back in the day understood much better than the present day. And even in countries like China, right, China has had explosive industrial growth and breakneck piracy figures. Like Chinese piracy and infringement is, they will find entire factories which are manufacturing stuff without a license from the IP owner. And China, I think, is very conscious they didn’t want to enforce IP rights too rigorously because it does impose a kind of tax and increases transaction costs for businesses. So back in the day in the 60s, sorry, in the 50s when we were looking at copyright law reform, the focus was to try and water down copyright law, reduce the term of protection, reduce the other kind of protection that authors were getting because policy makers at that point thought that our primary need today is access to education. So we need cheap textbooks, affordable textbooks, affordable learning material, we don’t want to spend money on foreign exchange, etcetera, etcetera. So that was the plan. It is a different thing that the authors’ lobby was so strong in India that they managed to stop it in parliament and you know by lobbying the education minister back then. But same thing with patents, right? It was the process started I think in 1948, finally by 1970 the government, Indira Gandhi’s government managed to enact a new patent law which basically said that ‘give up any pretenses of being like an innovator economy, etcetera, they said lets just copy and make and set up our industrial base.’ And it worked out pretty well for the pharma industry, there are quality concerns which are a different problem. But it has worked out very well in terms of, these companies did not have to secure licenses or they weren’t blocked from manufacturing new medicines, etcetera. So I mean, the Indian experience so far has always been that you don’t need strong IP for your economy to grow. It all depends at which stage of growth your economy is, and what a lot of policy makers, law students, law professors make the mistake of is imitating what is there in the West without always acknowledging the kind of differences that are there between the Indian economy and those other economies. Today, even if you say we have had a strong patent law, now quote unquote strong patent law for 20 years, we have not seen any great innovations coming out in India. And even in the pharma industry where there is expertise and there is access, they have the capital to invest in R&D right? Instead, it has acted as a tax on the Indian economy in the sense that the prices of some medicines have become more expensive and etcetera, etcetera. The reason that we still have it is because if you want to trade on the international scene, you want to sell your textiles to America, America is going to say you protect our IP, so that’s the reason we have it. And it is important to understand the dynamics of why we have IP. India has IP today because it is protecting pharmaceutical patents because it wants to trade on the international scene and countries in Europe and America, they want you to protect their competitive advantage which is in intellectual property and inventions. So all of this is, these are all nuanced questions, and it is important to remember that nuance when approaching it from a policy perspective, rather than just aping what is already there in developed countries which have a per capita income of, you know, $70,000, $80,000 per person. That’s a long answer to your question, but I hope I’ve answered it.

Shauryaveer Chaudhry: Yes, sir, thank you for that. You spoke about how imitation can itself be economically productive and disruptive. An example that comes to my mind is the claim of personality rights by celebrities against fan merchandising when these acts of merchandising itself is what kind of establishes their celebrification. Now, large AI models, coming to the topic of the symposium are trained on enormous quantities of existing human created material. You think that the instinct to treat copy during AI training as inherently problematic misunderstands the way that innovation kind of actually happens in the real world, or is it something else?

Prashant: I mean, see the AI issue is a factual question, right? Like even in this open AI litigation, the problem is we have not yet had a trial, so a lot of the facts are not yet clear. So for example, if it is a text and data mining style operation, you know where you are using a computer program to basically figure out statistical patterns like going through copyrighted work, but you are not reproducing anything that may not even count as copyright infringement. But, if you are doing wholesale copying of information, imitating it and putting it into your system, and then you are regurgitating the answers when a user makes a particular query that would be more of a copyright infringement case. Now, what is good, what is bad? these are all, I mean, these are value judgments. But if I was a policymaker looking at this, I would look at it from the perspective of what is the comparative advantage that my economy enjoys today, right? If my guys can’t develop an AI program they are unable to, they just don’t have the technical expertise or the whatever the risk capital to invest in developing an AI program and I have a huge creative industry which has, you know, books and music of value. Then logic would dictate that I protect the area in which I have a comparative advantage, which is the creative industry. So I will secure this and I will tell the AI guys, you have to pay for anything that you use of these people. The other way around to look at it is you want your AI industry to develop locally, you are willing to give them subsidies in the form of allowing them to skirt copyright laws to a certain extent, and you are not that worried about your creative industry because you know that no matter what this AI guy does, it is not really hurting their bottom line or their incentive to keep producing new music or new literature. So this all depends on how, as a policymaker, what are your priorities and at a deeper level, what is the political economy of your country. Back in the 50s imagine if a group of authors led by Ramdari Singh Dinkar could sabotage copyright reform, it just shows you how these politics can play out behind the scenes to affect how reform process work. Because I think back in the 50s we had a very good case for diluting copyright law as a developing country. And the copyright, I mean, the authors would have still made some money, maybe they wouldn’t have made the same kind of money, but it still got, you know, as a developing country with really poor literacy rates, we still managed to. The authors still managed to intervene and stop it. So these are questions where it’s difficult to give yes or no answers. I think it depends on these various economic factors and political economy of what is existing in society.

Shauryaveer Chaudhry: So just following up on that, knowing the ground, as you said, that we need to be cognizant of the ground realities of India. Knowing these ground realities as a policy maker, then which of the 2 approaches would you kind of feel is more suited for India? Because yes, we do have a very wide system of culture of creativity. But I do feel that at some point AI also is being given some kind of importance in the Indian technological industry as well. So as a policy maker who is kind of analysing for India, which of the 2 routes would you take? Would you take a kind of a mixed approach?

Prashant: I mean. I don’t think, I mean like right now the government has been talking about sovereign AI, etcetera, but apart from Sarvam AI, I don’t see any other effort ongoing to develop, you know, like an LLM or a large language model foundation in India. Am I mistaken on that?

Shauryaveer Chaudhry: It is an India AI mission, as far as I am aware, where there was a large corpus fund, it was given to kind of work towards the AI mission. There was something that was announced, if I am not wrong.

Prashant: Mhm. Yeah. I mean, so see if there is nothing happening, then you may as well just protect your guys against even the foreign large language model to make sure that at least they pay your publishers, etcetera, etcetera. Because one of the problems, one of the issues to remember is that, you know, when big tech initially took off 2025 years ago, talking about like, you know, Google, Facebook, etcetera. Nobody had an idea of how devastating they would be for journalism, right? Like they completely destroyed the ad market and that’s had repercussions down the line for journalism, etcetera, etcetera. So over here, it really depends on what your national priorities are. If AI is a game changer, I think in some areas, for example, like in a country like India for translations, AI can be a real game changer, right? In a multilingual country like ours. To that extent, yes, perhaps making some concessions for copyrighted works, etcetera, is a good idea, but if it comes to a scenario where copyrighted works are being reproduced and in a manner that their economic value falls in the future because an AI program is trained so well now on this that you don’t have to reach out for a book or a journal you’re disrupting a lot of, that entire publishing industry quite vigorously. So I would say, I would stick to the first principles of copyright law, which is if you are doing a text and data mining approach, if you are just like statistical analysis, you are not reproducing material in the whole, then I don’t think that accounts for copyright infringement. And, companies should be allowed to do that, but if these companies are reproducing books, music, wholesale and training their programs, that is reproducing, meaning the moment they make a copy on their server, that is reproduction. And however they use it after that, is up to them. But the moment that happens, I think the copyright owner deserves to get compensated to some extent. Now, whether that payment should be completely their monopolistic power or whether there should be some kind of, you know, state control on the pricing through our statutory license, single compulsory licensing model that I am open to. But I think there has to be some compensation because there will be serious downstream effects on the creative industries. And given what we have seen happen with journalism in the past, it will be irresponsible to not stop something like this happening in the future to the remaining public, the creative industries.

Aastha Nayak: So I think, thank you for that. I think this is a good segue to our next question. Another sort of core principle of copyright law is the idea-expression dichotomy. So do you think this is still a tool we can use for evaluating output generated by AI or does AI training and AI generated content as opposed to human generated content exposes sort of a weakness in the dichotomy itself that may be idea and expression are not as separable as courts have assumed previously. A case we found while researching for this was a US case of Bikram’s Yoga College versus Evolution Yoga, where a US court in California, I think held that a sequence of yoga poses was an unprotectable system and regardless of the creativity involved in arranging it. So the line that the judges drew there was that however originally arrangement is a functional sequence is an idea, not an expression. So generative AI models would complicate this line of judgment since they don’t copy expression directly, they use statistical patterns to generate new outputs that are consistent with existing patterns based on the data it has been trained on. So sorry for bringing it up with the context but our original question was, do you think that the idea-expression dichotomy is still relevant?

Prashant: Yeah, I mean, so I think, like I have kind of answered this already in the previous one, but… I see, I don’t think with AI it is that open and shut, right? I don’t think all of them are using a statistical inference model like what we would call text and data mining, right? Because I think with, for example, The New York Times case that is ongoing. The New York Times has evidence or apparently they are claiming that the OpenAI is reproducing chunks of their exact, verbatim reporting or their verbatim content. So if that is happening then that I think is open and shut copyright infringement depending on whatever fair dealing, American fair dealing exceptions. But otherwise, if it is just, you are just plotting out patterns, etcetera, etcetera, you could getaway. The question again I think is, is the reproduction being made on the servers of the company? So this is where it gets a bit more complicated, right? So for example, if it is all online content, like say, you know, ANI’s content here, if everything is available online, you let your program crawl through the page and make whatever statistical inferences, etcetera, that are required. You are not really copying anything onto your server. But with books, which is one of the main contentions over here with books, you have to convert the physical book into a digital copy. The moment you make that one copy itself, copyright law gets invoked. And then the question is, where is that copy being stored? Is it being stored on your server? Is there some 3rd server like there is one of these pilot sites, what’s it called? I think Annie or something, right?

Aastha Nayak: Yeah, Anna’s archive.

Prashant: Yes, Anna’s archive. So if this Anna’s archive which is hosting all this stuff where you don’t have to copy anything, but you’re just letting your program crawl through it and do the text and data mining stuff, you could, I think, get away with saying that there is no reproduction, so there is no infringement. So, but I mean just to clarify even your question a little bit, the idea-expression dichotomy comes in more at the stage of whether copyright exists in a material or not at the time of infringement.

Aastha Nayak: Right.

Prashant: So just to give you an example, I think there’s a famous Indian case on guidebooks for mathematics where, sorry, that’s the merger doctrine. That’s another one. Forget it. I can’t remember the case, but generally the idea-expression dichotomy comes up at the stage of whether ‘is there a copyright existing in a particular work?’ It doesn’t come in so much at the time of whether there’s infringement happening, right?

Aastha Nayak: So I have another question on sort of what we consider fundamentals of copyright law. So Section 57 of the Indian Copyright Act grants moral rights to authors. So in terms of AI generated content, who would you ascribe these modern moral rights to? Who would the author be?

Prashant: Yeah. That’s a good question. I mean, so I think there’s some case law which has said that AI cannot be an author, that it always has to be a human being. I think there is some American case law and English case law on the point, and which is, I think, the correct view point. Copyright law is always meant to serve as an incentive for human beings and a machine can never be an author. So the question then is, whether the person putting in the input into the machine to generate that specific output, can he claim to be the author? Because, for example, we have technically always allowed, say, photographers using machines to claim authorship over the final photograph or a video, etcetera. So I think that question perhaps could still be open. You could say that the user who’s generated, who’s put in prompts in a particular way to generate certain content can claim to be an author and then now associated moral rights with it. But I think it, I mean, it takes moral rights also into a dubious area, right? Because moral rights was is basically the right to integrity, the right, right to paternity, to identify yourself as this and the philosophical construct behind moral rights is this European approach to copyright law, where they don’t even call it copyright law, they call it author’s rights because they consider the work generated by an author to be an extension of the personality of the author. This is like going back to Kant and Hegel and all these German philosophers. So that’s why, like in Europe, copyright like authors’ rights is elevated almost to a human right. That a person is entitled to protect this because it is a part of his extension of his body. So, I mean, if, but if you are using any computer program like AI to intermediate this relationship and the output, then it is a good I think it is a good opportunity to relook even at moral rights and how it would operate in such a context.

Shauryaveer Chaudhry: Thank you for that, sir. Just one last question before we close today’s podcast. This book was published in 2017. Almost 10 years later, which argument in the book do you think has aged particularly well and which argument would you change if you were writing ‘Create, Copy, Disrupt’ today?

Prashant: Okay. The entire book has aged very well. I mean, see a lot of it is history. We weren’t very forward looking in the kind of content. I mean, we didn’t give any long term policy prescriptions. Sorry, now I’m just trying to recycle my brain through this book. I mean, because I mean, honestly, most of the stuff that we discussed, it ended over there. There hasn’t been too many, too much of like sequential stuff. I think like the traditional knowledge chapter, I think as you know, we have been particularly accurate on where this traditional knowledge digital library, we critique it, like Samiti and me, were big critics of it. I think we were the first people to really first or second perhaps, to critique it where the government is spending all this money on this light running this traditional knowledge digital library and the output is, I mean, what are you doing? You are protecting some random, you are challenging some random patents which have not much economic value and you are striking them down. And that continues to be relevant till day. That has not changed. Public money is still being burned. And the other issue I think, relevant is Basmati. We had an entire chapter on Basmati. 10 years later, the government is still losing cases because the government has basically taken on the responsibility of projecting Basmati. It is still losing cases across the world where Basmati now has pretty much become a generic mark and questions should be asked of, why didn’t we change strategy sometime earlier? And there are a lot of interesting questions on representation because geographical indication typically is to be administered by industry bodies. And so in India, if it is an agricultural commodity, farmers should have a representation and their questions of why didn’t we collaborate better with Pakistan because Pakistan also has a stake in basmati. So there are a whole bunch of issues over there which I think still remain alive. But I don’t think anything aged particularly badly in that. Sorry if it sounds smug.

Aastha Nayak: No, I don’t think it sounds smug. If you want to write a second part in like 5 to 10 years when we see how the questions, particularly the Delhi High Court is dealing with and eventually we will have more litigation on the intersection of AI and copyright? Would you want to write a second part?

Prashant: Yeah. Yeah, but I mean, not just from this AI, this Delhi High Court is, I am sure you guys will have the entire symposium, but I have been critical of the judgment already. It is unabashed judicial activism like I mean, Krishna Iyer would be embarrassed at how the judgment was written, right? But I think what is, or what would be more interesting is related to some of my doctoral work that I am doing now is plants and intellectual property. That’s an area that we completely missed, but it is a fascinating area on plant breeder rights especially. Because we stuck, I mean the only intersection our book had with agriculture was basmati, but there’s a whole world out there on plant breeders coming out with new varieties and getting IP protection, etcetera, etcetera. And then how do you protect even agricultural brands? I mean, champagne, at the end of the day, champagne, whiskey are all coming out of agriculture, right? Those models, yeah.

Shauryaveer Chaudhry: Our professor did discuss this, even when Aastha and I were learning intellectual property. Our professor did mention this, but this was more restricted to geographical indication more than anything else.

Prashant: Yeah, but even in geographical indications, the fact that in India the state has just taken over a lot of these marks is very, very problematic, right? Like in whether it is, for example, tea, Darjeeling tea is owned by the tea board. A lot of spices are owned by the spice board and these are all controlled by government bodies, whereas like whiskey in the UK. For Scotch you have distillers and then you have the guys who grow the Bell etcetera. Same thing in champagne, the wine owners and the bottlers are a part of the association with their own GI. So like in India, the fact that we don’t have these participatives or this thing where even workers get certain say at the table and the state is sitting over all of this like my above is very, very troubling. and I mean, that’s something I thought we should give a lot more attention to. We touched upon it when we did this, when we asked the question of why Basmati is being, you know, administered by IP. But I think, yeah, that’s another issue that I would look at in detail. Yeah. So, agriculture and IP I think would be the additional chapter. The book, I don’t think that much, because apart from this case, I don’t see much happening in India on open AI. Basically, if you look at the arc of intellectual property law, it is the jurisdictions where the technologies are developed first because courts and those jurisdictions that will set the law and everybody the market then adjusts itself to that. So whatever the American courts are going to do now over the next year or 2 will set what the standard market practices are going to be in India by the time our courts finish hearing these cases and coming out with judgments, it will be irrelevant is my guess.

Shauryaveer Chaudhry: Thank you so much for that, Sir. And this was a very lovely podcast. I am sure our viewers would really benefit from viewing this and reading the transcript as well. Thank you so much for agreeing to join us today as a part of the symposium. And Asthaa, do you want to add anything to this?

Aastha Nayak: No, nothing from my end. Thank you sir.

Prashant: My pleasure. Thanks for having me.