Aastha Nayak & Vinayak Kedia

Ujjwal Sharma and Vedika Mittal speak with LSPR Editors Aastha Nayak and Vinayak Kedia about the improvised architecture underneath India’s personality rights, a body of law built almost entirely out of ex parte interim orders against an anonymous “Ashok Kumar” defendant, never contested and never brought to final judgment. They discuss how a right assembled from Article 21, the tort of passing off, and a 2012 three-part test holds up once deepfakes, voice cloning, and posthumous digital resurrection are put to it; where the line falls between a celebrity’s commercial persona and an ordinary citizen’s claim to dignity and privacy; whether Indian courts have gone further than the UK’s more disciplined reliance on passing off, and at what cost to journalism, parody, and satire; and what an AI model’s use of a person’s likeness, at the training stage and the output stage, does to a copyright framework still built around the assumption that a human is doing the authoring.

LISTEN TO THE PODCAST

Aastha Nayak: Hello and welcome to the Law School Policy Review podcast. Our topic for today is the rise of personality rights in India and this episode is part of our mini symposium series on the role of artificial intelligence in society. Today, Vinayak and I, Aastha, will be speaking to Ujjwal Sharma and Vedika Mittal, from Sharma-Kemp Chambers. I will briefly introduce them for our listeners. Ujjwal litigates civil and public law matters before the Madhya Pradesh High Court and works on interim relief, service, and enforcement as matters of daily practice. He holds a BA LLB from Jindal Global Law School and an LLM from King’s College London. He is also one of the youngest people to argue before the National Green Tribunal. I think he did so at age 19. Our other guest is Vedika Mittal. She holds an LLM in intellectual property law from University College London and practices trademark and commercial IP law. We are really honoured to have you, sir, and ma’am, for this conversation today.

Ujjwal Sharma: Thank you so much, guys. Thank you so much. And we’re also looking forward to the interesting questions that you have ready for us.

Vinayak Kedia: Thank you so much for the introductions and for your wishes. Ujjwal sir and Vedika, ma’am, for some context in our topic, the Indian Parliament has never recognised personality rights, but still the Delhi High Court has proceeded to grant many injunctions since 2025 which cover voice cloning, face warping, chatbots and even avatars. The legal scaffolding for all of this is Article 21 of our constitution, the tort of passing off and a 3 element test from a 2012 case. All of it comes from one court and none of it is contested. The defendant is usually Ashok Kumar, which is the Indian version of John Doe. What we are calling a body of law is in fact a run of interim orders, one side of argument granted at speed. This episode examines this machinery of personality rights in India and how to improve it. We can start off with a question by Aastha.

Aastha Nayak: So my first question is for Vedika ma’am specifically. So since as an IP lawyer you must be spending a lot of your time on filing for trademarks and protection for brands. So can a person’s name or a personality become their trademark? I remember that in the 2012 Cyrus Mistry case Tata Sons sought to shut down websites that simply used the directors names. They claimed that this name itself was worthy of protection. So what happens when a celebrity or any other public personality walks in and wants to secure their brand using their name? How do they meet the criteria for trademark protection?

Vedika Mittal: When you talk about name or lets say protecting a face or a voice, I usually still start with the same IP toolkit, but then you have to stretch it a bit further. I would say when a conventional brand or normal business owner comes to us, it could be for a brand name, logo, tagline or a particular mark and I can build a fairly straightforward trademark strategy around it. But with working with personalities, sportsperson or where you could say the real commercial value is much broader, it could be their face on the billboard, their voice in an advertisement, a particular gesture, a catchphrase, or even a way of presenting themselves that the audience immediately associates with them. And that is where things become interesting, because in Indian trademark law, you will never see one neat category called a personality or a persona that I can register under. So as a lawyer, I have to then build protection through multiple rights. For example, I would look at the person’s name, signature, logo, or other distinctive identifiers from a trademark perspective. I would also maybe pay very close attention to endorsement agreements. Because the contract needs to clearly establish who can use the person’s image, voice, and other campaign material, where they can maybe use it and for how long they can use it. So copyright can also again become very relevant in protecting actual pictures, advertisements, or videos, and other creative material. When somebody goes ahead and uses the person’s identity without permission, we also look at passing off and personality or publicity rights. A very important case which I can think of is the Titan Industries case where the Delhi High Court recognised the right of publicity in the context of, I think, it was Amitabh Bachchan and Jaya Bachchan. They had essentially 2 important ideas over here: whether there is an enforceable right in a person’s identity and whether the person is identifiable from unauthorized use. So what I find particularly interesting is how the area has developed since then. A very recent case which I can think of is the Anil Kapoor case in which there was a mark registered to his celebrity voice, likeliness of his distinctiveness of his expression or even AI generated or deep fake content into the conversation. So in practice, I would say my toolkit is not one single law, but maybe a combination of trademark, contract, copyright, passing off and personality. That is why maybe I would say this area is actually becoming way more interesting because AI can maybe reproduce your face, your clone, your voice within seconds. And the question is not simply who owns the picture but who gets to control your identity. So those things I think are becoming very interesting.

Ujjwal Sharma: Yeah, I think I would like to say this without trying to offend Vedika or any of the other trademark practitioners. IP traditionally, for me at least in my experience, when I dealt with it, has always been very boring in terms of repetitive form filling, but I think this is something that is very interesting and is adding a bit of the juice.

Vedika Mittal: Yeah, it’s actually now becoming very juicy, and I see a lot of other litigators entering into IP.

Ujjwal Sharma: Yeah, definitely.

Vinayak Kedia: So as Ujjwal sir mentioned, this area is becoming very juicy and indeed it is exploding nowadays. But just to follow up, ma’am, a lot of the discussion that we have had relates to commercial protection of personality rights. So for example, Amitabh Bachchan or even the Anil Kapoor case, we are dealing with personalities who might want to safeguard their image so they can later contract or license it out. But what would you think about, for instance, deepfakes whose purpose is not really commercial gain, but actually just maligning the person involved? So, for example, in cases involving Taboo or Kareena Kapoor, we have recently seen sexually explicit deepfakes being floated on the internet. In this case, would the traditional IP route be helpful or would you make an article 21 argument there.

Ujjwal Sharma: Absolutely. Now that depends on 2 things, right? First is if lets say we are dealing with someone who has a personality, right, a commercial persona. And second, if let’s say we are dealing with regular people, or those who do not have a commercial persona around them. So now the strategy around those 2 changes very drastically, right? Obviously, when and this is very unfortunate to see that deepfakes and sexual content being circulated online and being published online. So then obviously you do not just stick to trademark law or IP law in particular. Why? Because then that becomes a broader issue of criminal law. Right, because their whole, it could have things to do with voyeurism, it could have things to do with defamation, right? And of course, privacy. So then you look at the more whole broader scheme of, let us say, article 21 and then the criminal side of things under the BNS. So then our strategy changes accordingly.

Vedika Mittal: I think I would just say that you do not have to be a famous personality to protect your own identity, privacy and dignity. So I think all of us as the citizens of this country under Article 21, have the right to dignity, we have the right to privacy. So that is, I feel is not that big of an IP issue if we were to understand the depth of when these things happen to, let us say, minors or like people who genuinely get affected by it. So, I think that is something we need to think about. And not just restrict ourselves to a commercial perspective only.

Aastha Nayak: I think it’s a really interesting answer because the right to privacy is usually under Article 21 after the Puttaswamy case, that is what the court has held. So if we are saying that personality rights emanate from broader fundamental rights, how do we look at the Ashok Kumar, which is, as Vinayak mentioned earlier, the Indian version of the John Doe order? How do we bind the defendant when there is no when it is an anonymous person? How under civil procedure, can we contemplate such orders? And as lawyers, who would you hold liable in these cases?

Vedika Mittal: Well, you know, I would say there is a lack of statute that we have in our country because you can’t… so I think by the latest development by the ministry of IT we do have this option where you can get the content removed within a particular period of time, they have to act immediately, but we still do not have these answers of who is behind this, who is uploading it, how many people are involved in it. So that is somewhere where I feel that there is a lack of legal statute in our country, but then again, a lot of times it is very easy to talk about laws, but when it comes to the implementation of it, becomes a very big challenge because it is easy to write, it is easy to talk, but then the ground reality is very different.

Ujjwal Sharma: So to add to that, let’s discuss what is and how do you actually serve an Ashok Kumar and what does the CPC actually permit, right? So CPC does not actually ever use the phrase John Doe, right? So what the courts have done is….

Aastha Nayak: Sorry, Ujjwal sir, you got cut off a little bit. We only heard that there is no term like Ashok Kumar or John mentioned in CPC. So what courts have done? Could you start from there, please?

Ujjwal Sharma: Right, so I’ll start again. Yeah. So now this is a very interesting problem when that comes with serving an Ashok Kumar and what does the CPC actually permit? So CPC does not actually use the phrase John Doe anywhere, but courts have derived that power from order 39 and its inherent jurisdiction under 151. The interesting problem here then comes in service. So order 5 broadly assumes that you can identify and locate a defendant, right? But what happens when you know your defendant is quite literally an Ashok Kumar from the Internet? So courts have dealt with that through substituted service including publications and then treating anyone who subsequently comes forward as being bound by those proceedings. So when we talk about personality rights, the Delhi High Court has gone a step further than this. So you often get a composite order which is a John Doe or an Ashok Kumar. This basically restrains from exploiting the plaintiffs persona. So now what happens is this is known for platforms as directing them to take down specified links and in the injunctions that can operate dynamically through allowing new links or URLs to be added through an affidavit rather than requiring a fresh suit to be filed every time. Now that is certainly at the outer edge of what the CPC had originally contemplated. But this isn’t entirely alien to what our Indian procedure has been. We have used John Doe orders for years in copyright privacy and you know, piracy matters and trademark counterfeiting matters. Now, this is essentially applying the same logic to someone’s persona. The real problem, therefore, isn’t so much how much the doctrine is practicable, but rather you have to see that you can bind parties within your jurisdiction, you can work with cooperative intermediaries, but try to serve an anonymous operator sitting anywhere abroad using a non compliant platform, which you know with how the Delhi High Court has taken a rather different proposition in this matter. So the court can definitely make the order. The Internet still has to cooperate with that, right?

Vinayak Kedia: That makes a lot of sense, sir. And obviously, you mentioned how the Delhi High Court issues these orders and it might be difficult to enforce them, but there’s another aspect of these orders which is also to me at least very interesting: they’re all interim orders that are granted ex parte, the other side, obviously, because they’re anonymous, they never appear. And they’re also interim in that final judgment is never awarded by the Delhi High Court in these cases. How do the two of you think this lopsided decision making affects even the development of personality rights as a doctrine in India? Because we have almost no jurisprudence which is finally decided.

Ujjwal Sharma: Absolutely. And I think there is a cost of building doctrines through interim orders which have no other contest to it, right? Because that is what is happening in this case, right? So there is a broader pattern here that isn’t unique to just personality right cases. We have seen it in rent control cases under a specific performance and now IP cases, of course, a series of unopposed interim orders, gradually they start to look like settled law simply because no one ever gets to a point of a fully recent final judgment. Now, if we look closely at the Sadhguru and the Variku orders, particularly about the court granting a very broad relief without always clarifying separate dignitary harm or consumer confusion or commercial misappropriation. And as someone who spends a fair amount of time in contesting interim hearings, that does concern me a bit because, see, when the defendant isn’t there, the court is naturally hearing one side’s best version of the case. And you end up developing very plaintiff friendly defaults in this case without properly testing the competing interest. Then of course, there is satire and journalism and the right to receive information and even the practical burden that these orders often place on platforms. And ultimately, that is what an appellate court is supposed to do. It is supposed to help us calibrate law. So the difficulty is that in most of these cases, they settle or remain at an ad interim stage and never get that full calibration that they might need. So you end up with an interesting situation here — Vedika and I were discussing previously when we were preparing for this podcast — that law appears to be developing very quickly, but the jurisprudence may actually be developing very slowly around this. There are plenty of orders that we have, we just don’t have enough litigation on those.

Vedika Mittal: See again, I think it comes back to the fact that we do not have a statute right now for these things. But again, what is the correct thing to include? How far can the parliament go to develop this and how effective will it be on ground? I think that is something that all of us keep thinking about and we keep, by each judgment, thinking of it as strengthening our position in protecting our clients, but again, the ground reality is very different and I don’t know how far can the parliament go to develop these statutes.

Ujjwal Sharma: Absolutely.

Aastha Nayak: And I think all the orders we have so far, at least on our preliminary research, are cases filed by celebrities or well known public figures. So since your chamber also does a lot of pro bono work, what if someone comes and they’ve had a deep fake or some sort of sexual content made of them without their consent? What can personality rights do for them? The route we discussed is if it is a commercial matter, but what if it is not someone who has a commercial interest in this?

Ujjwal Sharma: Right. So in that situation, those are very common, right? We have had cases in the past when we have dealt with AI generated content — and this is very unfortunate and very saddening again to see that there is AI generated or deep fake videos or materials of women being circulated online. Now you also have to think in all right with personality rights, that is a purely commercial angle, right? But with, let us say, someone who comes from a not so privileged background, right, someone who comes from a rural background.

Vedika Mittal: Also, I am sorry to interrupt, but like when you talk about someone with a personality again, I just feel that it is not fair to anyone. I don’t think it is fair to someone with a personality who has not given permission to use their content in that way. I don’t think it is fair to them as well. And I think in a matter like this, each individual should be treated equally.

Ujjwal Sharma: Absolutely, no, absolutely, that is there. So, but what I was trying to highlight is obviously, you see, if someone comes from a not privileged background or an underprivileged background and they reside, lets say, in a rural area. Now the thing is, often victims of these sorts of crimes are women who are quite young or who, lets say, lets assume a situation where a victim is a woman who is not yet married, right? And let’s say someone out of spite or how things very unfortunately happen in Madhya Pradesh — if there is a property dispute, they will end up filing a false sexual case against them, like a rape case against them or a POCSO matter against them. So if let us say something like that happens to a woman who is not yet married, you need to understand how difficult it will be for her to now get married because of obviously the taboo surrounded by this and all of the negative sort of image that it would create. So in this situation, I would not start with personality rights at all. If a person has no commercial persona, the strongest starting point would obviously be criminal law, privacy and platform liability. So on the criminal side, there is obviously the BNS that has provisions dealing with sexual harassment and voyeurism under section 77, which is particularly relevant for non consensual dissemination of intimate images. So this important point is that law is not limited to celebrities. The victims, as Vedika rightly pointed out, law is not just limited to or just for the convenience of celebrities. The victim does not need to have a brand value attached to their face. Now, then there is obviously also the IT framework that surrounds it. The intermediary rules require the platform to have a grievance redressal mechanism and respond to complaints involving privacy violation or sexually explicit material. So practically, one of the first things I would identify is every link that is there, every account that is there, and every platform that is involved and start pushing for takedowns through appropriate channels. And then there is also the constitutional dimension that is attached to it. As you guys have rightly pointed out, the Puttaswamy case makes it very clear that privacy and dignity are not just celebrity privileges. Each individual has a right to information, informational privacy, and bodily autonomy. So if someone’s face has been superimposed into a sexual material without their consent, the legal injury is not simply that someone has misused my image. It is more fundamental than that. If we talk about under the civil law, what remedies can you give them? I would therefore frame the case around privacy, dignity, and wherever it is appropriate, defamation, rather than just trying to manufacture a commercial personality rights, because that is not feasible in that matter. The immediate objective would be quite practical, that is to get the content down as soon as possible and preserve evidence that is there and identify the accounts and wherever possible, identify the person behind that. Because a victim in this situation — let’s say a beautifully worded judgment 6 months later — is not practically helpful for them. If let’s say a deep fake has been circulating through WhatsApp for 6 months, now priority is speed, evidence preservation and coordinated action under the civil and criminal regime.

Vedika Mittal: I think I 100% agree with you here and the most important distinction to make here: you do not have to be a celebrity to have a right to your identity, privacy and dignity. Lets say if someone came to me with a fabricated sexual video using their face, my first question will never be how commercially valuable this person’s image would be — it will be how do we stop the harm. I would rather straight up try to collect evidence, get this content taken down as fast as possible. The fact that the person has no commercial persona does not make the harm any less serious. In fact, in a case like this, the issue is not really about publicity or endorsement at all. It would be someone taking a person’s identity without their consent and attaching it to an intimate or sexual conduct that has actually never happened. Again, coming back to what Ujjwal said, that I would actually build my entire case around privacy, dignity, reputation, the criminal remedies, rather than just forcing the case into the personality rights framework. There is also a very practical side to this: when it comes to fabricated sexual material, speed matters enormously. I would want the URLs, the account details, the screenshots — as much evidence as I can gather at that point — and whatever is necessary to take down the relevant explicit content. And again, for me, if I had to put it very simply, I would just say personality rights protect the commercial and identifiable aspect of who you are, but privacy and dignity protect the person underneath that persona, which I think each individual deserves irrespective of what their personality is. And I feel it is something maybe you don’t need to have a brand, a brand endorsement or millions of followers to say this is my face, this is my identity, and I did not consent to something like this.

Aastha Nayak: As we were discussing earlier, how a dedicated statute for this — an amendment of the IT Act — would be helpful. So, do you think one statute can serve the interests of both these different sorts of classes of litigants, both celebrities and maybe people from rural areas?

Vedika Mittal: Again, when we talk about celebrities and when we are talking about something like statutes, I like to believe that all these statutes are made from the perspective of covering the benefit of as many citizens as possible. I do not think that there is ever a statute that is made for, let’s say, not even 0.5% of our population. So I do not think that there should be a statute particularly to protect the personality rights of celebrities. That is not something I would do right now. At this moment, personality rights in India are having a moment of their own, with the court decisions and how the Delhi High Court in particular is supporting and has responded quite quickly to the new problems. So that is also when we talk about personality rights — it is a very privileged and luxury part of law. Let’s not forget that. And then when we talk about… sorry, I think I skipped your question. Could you please repeat the last part of it?

Ujjwal Sharma: I think the question was, would a dedicated statute or amendment to the IT Act be beneficial, right?

Aastha Nayak: Yes, if you were an MP, what changes would you suggest for a law like this?

Vedika Mittal: So again, coming back to what I was saying, I would never want to frame a law that is for that privileged or luxury kind of people. I would rather want something that is for the masses, for everyone, that would maybe give more protection to people at large. What exactly is the protected attribute of a person’s persona? Is it only their name, only their image, or would you go to the extent of protecting their voice, mannerism, catchphrase and other distinctive characteristics? Is that the right commercial way to go ahead? What happens when the person dies? There are a lot of questions to think about, because again, we don’t have a statute and we are just trying to have a conversation and think about what would be the correct way to go about it. So personally, I would be open to a narrowly drafted personality rights statute that gives us some basic rules without creating an unlimited proprietary right over someone’s dignity. It could maybe define what kinds of attributes can receive protection, identify the circumstances in which unauthorised commercial exploitation becomes actionable, provide a separate and carefully defined remedy for serious invasions of privacy or dignity, which is very, very important, and clarify the questions of duration, inheritance, and posthumous rights. But I would be equally concerned about what that statute should not do. I do not think that it should give monopoly to the privileged over everything that they own — that should not happen.

Ujjwal Sharma: Everything that references them, shouldn’t be protected.

Vedika Mittal: Yeah, maybe we need better safeguards for journalism, criticism, parody, artistic expression, and whatever is legitimate for public interest. I would never want to replace the existing judgment law entirely — I would maybe use the legislation to give it a clear framework.

Ujjwal Sharma: Sorry to interrupt you. So what came to my head — if we give a broader sort of monopoly to celebrities over every reference that is made to them, then we would also be stopping what was originally seen as one of the most important aspects of journalism, that is, the cartoons that they used to draw. So if you think from that perspective — and it is very sad to see the decline of that happening in India now, but earlier and elsewhere as well — these cartoons have always played a very important role in telling the population as to how their government is functioning or what is happening. Or even in the newspaper, the section used to be an attractive bit for kids too — even though I would not understand most of it, I would love to see those as a kid.

Aastha Nayak: I think that we need to balance the right of free speech of citizens and the media. I think right now a lot of the courts tend to favour the commercial benefit of not just celebrities, even politicians in the current landscape.

Ujjwal Sharma: Absolutely. Yeah, absolutely. And secondly, also, another thing that I can now think of from this — do you remember the movie, there is a horror movie called Scream, where there is this guy with a very elongated mouth?

Aastha Nayak: Yes. Okay.

Ujjwal Sharma: There were a couple of movies, parodies of horror movies, which are comedy horrors, called Scary Movie. There is a 5 movie series part of it, and there is a very famous catchphrase from it, a very famous GIF that I still send, which I don’t think you guys, genz, would appreciate much. There was a very famous GIF of “what’s up” where the guy sticks out his tongue. So if we give celebrities a monopoly over every reference of themselves, we will be stopping these artistic expressions or parodies that are created, right?

Vedika Mittal: My second point, coming back to the question, would be that we also have a lot of AI generated content. And so now the government has started — they have asked for all the synthetic content that is created to be labelled, traced, and removed. And here, interestingly, Parliament and the Ministry of IT have already, quite substantially through 2026, made a lot of amendments to IT rules. The amendment framework now specifically addresses synthetically generated information and imposes technical and due diligence obligations on the intermediaries, including measures to prevent unlawful synthetic content and requirements for labelling it. So rather than maybe creating a completely different AI statute just to deal with deep fakes, I think what I would, in my humble opinion, suggest would be 2 complementary layers: clear substantive rights for individuals and stronger platform obligations for the technological ecosystem through IT rules. Because I think one of the biggest problems is the fact that the minute someone generates this content, it can be circulated within minutes. So I think something like this would address both sides of the problem. One tells the victim what right has been violated and the other tells the platform what it has to do about that content.

Vinayak Kedia: I think that’s a very interesting framework of making 2 layered individual rights and then obligations for the platforms. But, and like you also mentioned, ma’am, individual rights, we have to define them in the statute and perhaps that is the role it can take. So if you could take this conversation to another step — in the recent Sushant Singh Rajput case, the Delhi High Court ruled that personality rights are personal and they’re not heritable. So can they completely extinguish after a person dies? And even Sir mentioned how you want to safeguard and balance the individual personality rights and the broader rights of others. What do you think is a better approach in today’s age? Should we protect the posthumous personality rights of a celebrity, or should we say that they die with their death and anybody can make whatever content they want after their death?

Vedika Mittal: I think this is actually one of the places where AI has exposed a very real gap in our existing laws. Let’s assume very unfortunately, if someone dies today, their voice does not necessarily disappear from the commercial world, right? We still have years of recordings, interviews, films, songs, and all kinds of generative AI that we can potentially use to synthesise and maybe make them say something they’ve actually never said. So the question becomes, who controls that voice after the person is gone? India does not have a clear statutory answer to that question as of now. I think one of the very important cases would be the Krishna Kishore Singh case, in which the father of, as you just mentioned, Sushant Singh Rajput wanted to seek restraint on films and other work based on his son’s life. The Delhi High Court held that privacy, publicity, and personality rights being asserted on behalf of Sushant Singh Rajput were not heritable and had ceased with his death. The court also relied on the principle that certain purely personal causes of action do not survive death, while recognising that the broader question of the nature and survivability of commercial celebrity rights may require deeper examination. So as things stand, I would be very cautious telling a celebrity’s family that they automatically inherit broad control of that person’s rights after their death. At the same time, I think AI voice cloning makes that policy problem much more difficult. When this case was decided, there was an important distinction between making a biographical film about a deceased actor based on facts already in the public domain, and somebody commercially cloning an actor’s voice and making it appear that the actor is endorsing a product or speaking words they never spoke. The first raises an obvious question about freedom of expression and the public’s ability to engage with the life of a public figure. The second is much more like commercial appropriation of an asset that the person had built over their entire lifetime. And then this is where I think legislation could be useful, if we decide some personality or publicity rights should survive death. The law would need to tell us who owns them, how long they survive, whether they can be licensed, whether they can be inherited, and most importantly, what happens when the use is expressive rather than commercial. For example, should a family have the right to stop a documentary? Probably not, simply because it uses the deceased person’s name or voice. But should a company be able to create an AI generated voice of that person and use it to sell a product without anyone’s consent? That is a much harder case to justify. So I think AI has actually changed the urgency of the debate. Earlier, posthumous personality rights could seem like a relatively narrow question about films, pictures, and merchandise. But now, thanks to AI, you can reproduce indefinitely even after death. And unless the law gives us some clear boundaries, we are going to keep asking the courts to solve essentially a legislative question.

Ujjwal Sharma: There’s one thing that I wanted to add to that. So once you add indefinite voice synthesis and hyper realistic video defects into this equation, I think the problem becomes much more interesting. The same technology that allows a studio to recreate a deceased singer’s voice for an unauthorised tribute can now also put that voice into political propaganda and obscene content decades after that person has passed away. And that is where I think law should start asking some very uncomfortable questions. Most of the recent Delhi personality right cases, be it Amitabh Bachchan, Anil Kapoor, or the cases we have seen with cricketers and politicians, concern people who are very much alive and are able to walk into the courts themselves. We haven’t really had a definitive case on whether an estate can control someone’s digital resurrection after their death, and the courts don’t have a particularly neat answer to this either. Puttaswamy gives us a very strong conception of privacy and dignity during life, but doesn’t establish any clear rights over privacy after death. Succession law raises certain questions of which personality rights survive after death. And cases such as Sushant Singh Rajput also show that judicial caution about allowing publicity claims to interfere with expressive works is important. So where does that leave us with a synthesised voice? If, let’s say, a company uses a deceased singer’s voice to sell a product without their permission, as ma’am has already clarified, there is a fairly intuitive argument that this is exploiting commercial value created by a person during their lifetime, right? You can see why an estate would say that value doesn’t suddenly disappear when that person dies. But an intuition isn’t the same thing as a legal right. Without legislation, a court would only be able to go so far with questions like: do personality rights survive death, who controls them and for how long? AI makes that question much more urgent because unlike a photograph or an old recording, a synthesised voice does not simply preserve someone’s identity — it keeps producing new versions of that identity indefinitely. So if India does eventually legislate on personality rights, I think rights after death cannot be an afterthought. The statutes will have to answer the rather uncomfortable question: does the right survive after death? Who inherits it? And when does it expire? Because none of these rights can go on into perpetuity. Or can someone’s digital ghost live forever? I think that is a question that we need to ask.

Ujjwal Sharma: There is a very fun trivia that I would like to share with you too, if you ever go to London on the central line. What happens is, in London, on all the tube stations there is a recording playing in the background of “mind the gap.” Now, these were done quite when the tubes had started originally as well, a lot of these. So there is one station where a voice of a man was heard playing “mind the gap.” Now what happened is, after that man’s death, his wife did not have any other memories of that person that she could save apart from all of this, and she wanted to save every piece of him that was alive. So she requested TfL to keep one recording of this person on the tube station that still plays. So even now, if you go to — I think that is Baker Street station, I can’t remember correctly — but one of the stations in central London, if you go there, it will still play that original recording, recorded quite a few years and decades back, saying “please mind the gap.”

Aastha Nayak: So actually I have a slight follow up question on this. So when we studied intellectual property law and we were looking at UK cases, they don’t really recognise publicity rights, from our understanding. They use the tort of passing off — they rely on that more. So since you both work in India and studied law, and Vedika ma’am’s LLM was in IP law, do you think that India’s way of handling it through personality rights is better, or the UK method of relying on the tort of passing off is better?

Ujjwal Sharma: See, there is no one answer for this. I would not just say that one system is simply better than the other. I think they reflect 2 completely different approaches.

Vedika Mittal: So the UK is generally more cautious about creating a standalone proprietary right in a person’s image or persona. Celebrities have had to work through an established cause of action, particularly passing off, to establish their rights. One very famous case which I can think of right now is Rihanna versus Topshop, in which the courts were looking for traditional ingredients such as goodwill, misrepresentation, and damage. Whereas Indian courts, particularly the Delhi High Court, have been much more willing to expressly talk about personality and publicity rights and connect them with privacy and dignity. In India, this has become particularly significant with digital content. The Delhi High Court has now granted protection in cases involving not just photographs or merchandise but also voice, likeness, and AI generated or deep fake content — there is a 2026 order for a cricketer, I can’t recall the name, I think that was Gautam Gambhir’s case, where the court restrained unauthorised use of his name, image, and voice. So there is a clear advantage to the Indian approach — it gives celebrities a potentially faster and more direct remedy against forms of digital exploitation, which traditional trademark law was never really designed to handle. But there are also, I will say, maybe a few concerns here. If a personality right is framed too broadly, particularly through an ex parte injunction, there can be tension with freedom of expression, parody, criticism, and journalism. And that is probably the challenge for Indian courts going forward — how do we protect the commercial value and dignity associated with someone’s identity without turning personality rights into a monopoly over everything that can be said, shown or created about a famous person? I think that balance is going to become one of the most important questions in the Indian personality rights scenario.

Ujjwal Sharma: And just to pitch in from a litigator’s perspective, having litigated in both the UK and in India — from a litigator’s vantage point outside of Delhi, I would say that India is currently more permissive in terms of the outcome, but perhaps not quite as disciplined in terms of the legal framework behind that outcome, right? So the Supreme Court judgment in Puttaswamy gave us a very rich understanding of privacy and autonomy, as I have already mentioned a couple of times now, but importantly, it also gives control over personality. But once you have that constitutional language available, it is perhaps not surprising for litigants to start using it to build personality rights claims. But there is a very interesting tension here — in cases like Krishna Kishore Singh, the Delhi High Court was very careful to point out that Parliament had not actually enacted a dedicated publicity rights regime. So in that sense, we have courts being asked to build a fairly modern right using legal building blocks that were designed for a somewhat different purpose. And this is where I think the UK comparison becomes interesting, as the UK has generally been very reluctant to recognise standalone publicity rights. You can’t just simply walk into court and say “this is my face, this is my name, this is my persona and therefore I own the commercial rights to it.” Law there tends to rely more thoroughly on existing causes of action, particularly passing off and confidentiality and related doctrines. India, by comparison, I think sometimes seems to be saying, “well, we don’t have a statute for this yet, but we do have privacy, dignity, personality, so let us see how far we can take those.” And I think this works reasonably well when dealing with fairly obvious cases. The difficulty, I think, lies in the ones in the middle. What happens when there is a parody? What happens when there is a biographer? What happens in the case of comedians and journalists, and at what point does protecting someone’s personality become protecting them from being talked about? Because the danger is that we might end up in a situation which is very powerful, developed from one interim injunction at a time, which is rather an unusual way of writing legislation. Again, as we talked about at the very beginning, law should not be developed through interim injunctions, and as AI makes it increasingly easy to reproduce someone’s voice, I think those borderline questions are not going to become less frequent — they are going to be the main event of any case. So I wouldn’t necessarily say that India’s approach is wrong, but it may need to be more responsive to the realities of technology. I do think the next stage has to be about bringing some discipline and predictability into the law, because law is only as good, at least in a common law jurisdiction, as it is predictable.

Vedika Mittal: I think India would only be uptight about the personalities of comedians. That is where I think the UK is uptight at all times, and India is uptight when it comes to comedians.

Ujjwal Sharma: That is, unfortunately, something that we have seen. But without going further into that rabbit hole of things.

Vinayak Kedia: Right, very interesting discussion, and as you guys have mentioned, it is very difficult to define where you would place personality rights on the spectrum. Aastha and I were thinking, if you are talking about posthumous rights, for example in the Krishna Kishore Singh case, you could technically argue that even Ashoka the Great from 2000 years ago should have personality rights, Mahatma Gandhi should have personality rights, and at that point — where are the limits of the discussion and who has what rights? So certainly I agree that statutes should be there to define what these rights are. We have discussed a lot on this, but if you could move to another thread regarding actually practically securing personality rights — both of you deal with IP litigation, and Ujjwal sir, you’re also a certified mediator and arbitrator. How would you think ADR, or alternate dispute resolution, can be used when dealing with disputes over personality rights or publicity rights? Do you think it is a better solution to try to mediate or negotiate these rights, or is it simply that you get injunctive relief from courts where you have no other guidelines?

Ujjwal Sharma: Right, that is a very interesting question. I think ADR has a role here, but probably not in the first 5 minutes when something like this goes viral, right? Because let us say we are talking about sexual content here, sexual material being circulated. If someone’s face has been put on that and it is spreading rapidly, you need to go to a court and get a takedown order, and wherever necessary, you need to take criminal or regulatory action. You cannot mediate your way out of a viral deep fake. So the Delhi High Court’s recent approach reflects this urgency of ex parte injunctions and dynamic links being circulated, with tight timelines for platform compliance. The idea is fairly simple: first you stop the bleeding, and then you discuss the dispute. Where I think arbitration and mediation become useful is the second layer — after you have stopped the bleeding. For example, let us say there is a celebrity and a brand in dispute over an unauthorised endorsement — the scope of an existing license to use someone’s image in a campaign, or the revenue sharing around merchandise. Those are fundamentally commercial disputes; you are arguing only about money, about contractual rights and future relationships — this is where ADR can be extremely effective. And there is also an important question of arbitrability. You cannot privately arbitrate away statutory obligations or constitutional rights. You have to think about what disputes are arbitrable and what are not. Platform obligations under the IT rules, or criminal liability under the BNS or IT Act, or an individual’s fundamental right to privacy — those matters are not between 2 private parties arising out of a contract. So I don’t think ADR would be an effective mechanism to resolve those sorts of problems. I would see ADR as the second layer of response, not an emergency response — first you stop the content, preserve the evidence, then deal with the commercial or contractual fallout through the most effective route available. And sometimes, in a dispute involving a celebrity and a brand, the best outcome, honestly, isn’t a 300 page judgment — it’s getting both sides into a room and making sure that they never need to create another deepfake of each other again. That’s how I would summarise it.

Aastha Nayak: So, Vinayak, do you have a follow up or should I ask the next question?

Vinayak Kedia: No, I think we can move on.

Aastha Nayak: Yes, so with AI generated deepfakes and AI generated content in general — it would be trained on thousands of publicly available interviews, photographs, speeches, and other performances of the celebrity that are available online. So at what point do we say that some amount of that is okay for training AI, given how AI and large language models are trained? Is it an infringement of the personality or publicity rights of a celebrity? Where would we draw the line? From my understanding, section 52 of the Copyright Act, which talks about exceptions to infringement, would be applicable — but would Indian law permit the use of a celebrity’s likeness for training AI?

Vedika Mittal: I would again separate it into 2 stages: what the AI system ingests and what it ultimately produces — these are 2 completely separate things. At the training stage, the legal position in India is genuinely unsettled. Our Copyright Act doesn’t contain a specific text and data mining exception, unlike some jurisdictions that have created statutory TDM frameworks. So if an AI company scrapes thousands of interviews, photographs, performances, and books to train commercial models, the obvious question is: which section 52 exceptions are you relying on? Some argue that training is fundamentally different from copying for human consumption — the material is processed into parameters rather than simply reproduced and supplied to users. They argue this could potentially fall within research or fair use type exceptions. But there is a fairly obvious counter argument: section 52 contains specific enumerated exceptions, and commercial scale scraping to build a proprietary AI model doesn’t fit comfortably with many of them. The ANI media litigation is particularly important here because it puts this question before the Delhi High Court, although we do not have a definitive answer yet. The output stage, however, is a very different story. If the model starts generating content that convincingly sounds like a particular celebrity, looks like them, or impersonates their mannerisms, particularly for commercial purposes, you are much closer to the personality rights cases we have been discussing — Anil Kapoor, DM Entertainment, and other famous celebrity cases. This gives courts a doctrinal vocabulary for dealing with unauthorised commercial exploitation of a recognisable persona. So I would draw the distinction this way: the training question asks, can you lawfully learn from this material? The output question asks, what are you doing with what you’ve learned? And those are not necessarily the same legal questions. A model being trained on a photograph doesn’t automatically mean every subsequent output is unlawful. Equally, the fact that a model can reproduce someone’s style doesn’t necessarily mean every stylistic imitation infringes a personality right. The difficult case is exactly where that boundary sits for both. On section 52, I think the honest answer is that we do not have enough certainty. If India intends to permit large scale commercial AI training, it may ultimately need to say so expressly, rather than asking courts to stretch exceptions designed for photocopies and libraries to deal with trillion parameter models. That’s quite a technological leap for section 52 to make on its own — I don’t think anyone was thinking of this when it was drafted.

Vinayak Kedia: So just a small follow up, Vedika ma’am — we have discussed the training stage and the output stage. On the training stage, you’ve made clear that we need more law to clarify things. On the output stage, you mentioned it might relate to the traditional personality rights doctrine we’ve discussed so far — but then how would you place responsibility? Would it be the AI developer, like OpenAI or Anthropic? Would it be the person who prompts the AI to generate the likeness of a celebrity? Or would it be the person who posts that AI generated deep fake on the internet?

Vedika Mittal: I feel there isn’t one single person who can take complete liability for it. I think there needs to be a joint or dual based system, where there is a division of responsibility and a proper framework with specific penalties if those boundaries are crossed. I think this has to be a very carefully thought through statute, which Parliament will have to come up with, given the dynamic times we live in. To answer your question, we will have to pose another question first: do we need legislative intervention, or can existing principles of authorship and originality accommodate this? I think that is what we will keep coming back to. Indian copyright law is still built around the fairly simple assumption that there is a human author somewhere in the room. The Copyright Act still defines authorship in human terms, and the Supreme Court’s originality test, going back to the Eastern Book Company and D.B. Modak case, looks for intellectual effort and a modicum of creativity. So when an AI is merely assisting a human, the existing framework can probably accommodate that. But the difficulty begins when the human contribution is essentially just “write me a song about heartbreak” and the machine does everything else. At that point, who is the author? I think there are broadly 3 possibilities. One is to say the output isn’t protected because there isn’t sufficient human creativity. The second is to recognise the person who exercised meaningful creative control — perhaps through detailed prompting, selecting, editing and curating — as the author. And the third, more radical option, would be to create some kind of separate right for AI generated outputs without pretending that AI itself is the author. From a litigator’s perspective, I suspect courts are going to be very pragmatic before they become philosophical about this. The dispute we are likely to see first isn’t going to be “your Lordship, who is the true author of this AI poem” — it’s going to be more like “this AI generated poem or song sounds suspiciously like mine, is that infringing my copyright,” or “if the AI generated advertisement looks like my campaign, is someone passing off my goodwill or my work.” So initially, courts can simply apply familiar concepts about substantial similarity, originality, and market substitution rather than trying to reinvent authorship overnight. But there is a real limit to how far existing doctrines can stretch — if a human gives a prompt to an AI, and the AI independently determines the words, composition, visual elements, and melodies, it becomes difficult to say the human has exercised the kind of creative control that traditional copyright doctrine requires. I studied some of this at the University of Oxford, with my professor, Dev Gangjee — he’s a brilliant person in this field and has done quite a lot of extensive research that you guys should read. I remember asking him about DALL·E, since that was the hype at the time — what happens if I prompt it to generate an image of a panda sitting in a pool that’s on the moon? Is that my creativity, or DALL·E’s imagination? I think human-AI collaboration could probably fit within existing law where there is genuine human creative input. But for a genuinely autonomous AI output, I think we probably need Parliament to make a choice.

Vedika Mittal: Because I feel with our existing laws, if there is a substantive human element that has gone into creating the content, we already have laws for that, which have been around for years. But if, as Ujjwal was just saying, there is very minimal human effort involved, then I don’t know what you feel.

Ujjwal Sharma: So I suspect the first question Parliament will have to answer is not whether AI can be an author — it will be, if AI isn’t the author, does anyone else own the copyright at all? Because sometimes the most legally important person in the room is the one who isn’t there. So I think that is how I would answer that.

Aastha Nayak: I think that was a really interesting answer, and given the rate at which this field is developing, I think we could have another hour long conversation just on AI, but unfortunately we are slightly short on time. But an interesting question I don’t want to skip — Sharma-Kemp Chambers runs a lot of workshops for students and young lawyers, and Vinayak and I are also in our second and third years of law school. So when we are reading these Delhi High Court interim orders, you can look at it from different lenses — a CPC lens on interim orders, or an IP framework. So when we are reading these orders, first, what should we keep in mind, and second, over the next 2-3 years, what do you think the progression of these will be? Because eventually we will have to have final judgments, or the Supreme Court will have to play some role.

Ujjwal Sharma: So the first thing I tell law students is: look very carefully at what the procedural posture is. An order like Anil Kapoor or Sadhguru can look extremely powerful when you first read it, but remember, it is an ex parte ad interim order, often made on a very compressed record. So don’t read it as though you have just discovered the next landmark constitutional judgment — read it as a snapshot of how a judge is responding to a problem at an interim stage. That’s important, but you need to remember it is not necessarily settled doctrine. Second, separate the different legal strands. When you read a case like Titan, ask yourself: is the court talking about passing off and goodwill? Is it talking about privacy and dignity? Is it actually recognising something closer to a standalone publicity right? That distinction becomes particularly important in the newer Delhi cases, because those concepts sometimes start appearing in the same judgment. Do you want to take this forward?

Vedika Mittal: Yeah, I think Article 21 privacy and dignity come from the very famous case we’ve discussed a number of times, and not from whichever Delhi High Court order you happen to be reading that afternoon. Criminal liability in deepfake and image based sexual abuse cases comes from the BNS and the statutes, not from these commercial cases. So maybe my advice to young lawyers is quite simple: don’t just read the operative portion of the order or copy the prayer clause. Ask questions when you are reading it — what was the court deciding, at what procedural stage was it, what legal proposition does the judgment actually establish? This habit will save you from citing an interim order as if it were the constitution itself, which unfortunately is a mistake a lot of young lawyers are rather good at making, even in the moot courts we have seen so far. And I would also add one more discipline — whenever you get excited about a new personality right order, go back to the case that builds the underlying building blocks — for Titan, publicity rights and misrepresentation; the famous Shivaji Rao Gaikwad case for regional film industry persona; Sushant Singh Rajput’s case for posthumous rights. Then read at least one serious critique of these orders, which would help you before you treat it as a model. So always read everything critically, as a lawyer.

Vinayak Kedia: Thank you so much for that feedback, which we will always keep in mind, and for this conversation — Vedika ma’am, I feel this has been a very enriching discussion for both Aastha and me, and I am sure our audience will enjoy it as well. Thank you so much.

Aastha Nayak: Thank you.