Technology has always unsettled the creative economy and the law has always chased to catch up. But no technological development in the history of the entertainment industry has created as direct, as immediate and as commercially devastating a threat to individual performers as generative artificial intelligence voice cloning and deepfake video synthesis. These technologies do something that no previous form of imitation or performance art could do: they replicate a performer’s identity at scale, with high fidelity, at negligible cost, without the performer’s knowledge and as the Indian courts have now confirmed across a series of rapidly accumulating cases without any legal authorisation from the performer whose creative personality is being commercially mined. A singing voice that took a performer decades of practice and artistry to develop can now be cloned from a few seconds of audio and deployed in unlimited synthetic recordings. A film actor’s face and mannerisms can be transposed onto any video content, used in fraudulent endorsements, placed in pornographic material or embedded in political disinformation, by anyone with a consumer-grade personal computer and access to freely available AI tools.
India’s legal response to this challenge has been rapid and judicially creative, but it is structurally incomplete. The reason for that incompleteness lies in the architecture of the Copyright Act, 1957 and specifically in the provisions governing performers’ rights under Sections 38, 38A and 38B of that Act. These provisions, substantially overhauled by the Copyright (Amendment) Act, 2012, in partial compliance with India’s obligations under the WIPO Performances and Phonograms Treaty (WPPT), grant performers a carefully defined set of economic and moral rights in their live performances. They do not protect a performer’s voice as such. They do not protect a performer’s face, appearance or visual identity. They do not address the creation of synthetic AI-generated replicas of any of these attributes. The gap between what performers’ rights law protects and what AI voice cloning and deepfake technology can do to a performer is, accordingly, substantial. Indian courts have responded to this gap by drawing on constitutional privacy doctrine, the common law tort of passing off and the judge-made right of publicity developed through two decades of celebrity personality rights litigation, to fashion interim injunctive relief that functions as a de facto protection against AI-driven identity exploitation. That judicial response is this article’s central subject its statutory foundations, its leading cases, the doctrinal questions it leaves open and the case for legislative reform that the current state of the law urgently presents.
The Foundational Statutory Framework Performers’ Rights Under the Copyright Act, 1957
The Indian Copyright Act, 1957, first recognised performers’ rights with the Copyright (Amendment) Act, 1994, which inserted Section 38 into the statute. Before that amendment, there was no performer’s right of any kind in Indian copyright law. The practical cost of this absence was demonstrated vividly by the Bombay High Court in Fortune Films International v. Dev Anand, AIR 1979 Bom 17, in which the legendary actor Dev Anand sued to prevent the commercial exploitation of films featuring his performances in a manner he had not authorised. The Bombay High Court held, with regret, that there was no legal right available to a performer under the Copyright Act as it then stood to control the use of their recorded performance the copyright in the film vested in the producer and the performer had no separate enforceable right in their own performance. That ruling exposed a structural vacuum that took fifteen years and a specific statutory amendment to correct.
Section 38 of the Copyright Act, 1957, as it stands today following the 2012 amendments, grants every performer a “special right” called the performer’s right in any live performance. The performer’s right subsists for fifty years from the beginning of the calendar year following the year in which the performance was made. The provision defines “performer” by reference to the inclusive definition in Section 2(qq) of the Act, which encompasses actors, singers, musicians, dancers, acrobats, jugglers, conjurers, snake charmers, persons delivering lectures and any other person who makes a performance. This deliberately wide definition ensures that the full range of live artistic expression falls within the performers’ rights framework it is not restricted to mainstream film or music performers but extends to classical musicians, folk performers, stand-up comedians and others engaged in live performance.
The 2012 amendment introduced two further provisions Sections 38A and 38B that together constitute the substantive core of India’s performers’ rights regime and represent a significant legislative improvement over the pre-2012 position. Section 38A confers exclusive economic rights on performers. It grants the performer the exclusive right to make a sound recording or visual recording of the performance, including the right to communicate the performance to the public; to reproduce the sound recording or visual recording of the performance; to broadcast the performance; and to sell or give on hire or offer for sale or hire, any copy of the sound or visual recording of the performance. A critical proviso in Section 38A(2) addresses the position of performers in the film industry specifically: where a performer has, by written agreement, consented to the incorporation of their performance in a cinematograph film, neither the performer nor any person authorised by the performer can object to the enjoyment of the film by the producer. However, the performer retains the right to equitable remuneration from the producer if the sound recording of the performance is used for any purpose other than in the film a provision that protects the performer’s royalty interest in the commercial exploitation of their recorded performance beyond the film itself. Section 38B confers moral rights on performers: the right to be identified as the performer of the performance except where omission of credit is dictated by the manner of the use and the right to restrain or claim damages for any distortion, mutilation or modification of the performance that would be prejudicial to the performer’s reputation. These moral rights subsist independently of the economic rights and cannot be assigned, though they can be waived in writing.
What Performers’ Rights Do Not Protect The Voice Cloning and Deepfake Gap
Understanding what Sections 38, 38A and 38B actually protect requires understanding, with equal precision, what they do not protect because the boundary between the statutory protection and the statutory vacuum is the exact space into which AI voice cloning and deepfake technology has moved.
The performers’ rights provisions protect a specific, identifiable performance. Section 38 grants a right in a “live performance.” Section 38A’s economic rights apply to the making of “a sound recording or visual recording of the performance.” These formulations are tied to a discrete event of performance a concert, a film appearance, a studio recording session. They are not formulated as rights in the performer’s voice, appearance or identity as general attributes. A performer’s right protects the recording of the specific performance, not the voice that was used in performing it. This distinction is not merely technical it has concrete and serious consequences. When an AI system clones Arijit Singh’s voice using audio samples from his existing recordings and generates entirely new synthetic recordings of him performing songs he never actually performed, those synthetic recordings do not reproduce any specific protected performance. They replicate the voice as an attribute, the vocal style, the timbre, the characteristic phrasing. The performers’ rights provisions of the Copyright Act, 1957, as currently drafted, do not clearly reach this conduct. The sound recording of the original performance from which the AI was trained may attract copyright protection under Section 13(1)(b) of the Act, but the synthetic output the AI-generated voice clone is a new recording that exists independently of any specific protected performance.
The deepfake gap operates similarly on the visual side. Section 38A grants a performer the right to control the making of a “visual recording” of their performance. A deepfake video is not a recording of a performer’s performance it is a synthetically generated video that superimposes a performer’s likeness onto an entirely different person’s recorded performance or onto entirely generated content. No Section 38A right is engaged by its creation. Section 38B’s moral right against distortion or mutilation of a performance requires an actual performance to be distorted or mutilated a deepfake that manufactures an entirely new performance-like appearance of a performer, rather than modifying an existing performance, arguably falls outside even this provision.
It is this gap that explains why the Indian courts, in every one of the landmark AI voice cloning and deepfake cases decided since 2022, have grounded the relief they granted not in Sections 38, 38A or 38B of the Copyright Act, but in the constitutional right to privacy under Article 21, the common law tort of passing off and the judge-made right of publicity or personality rights built up through the pre-AI-era celebrity merchandising case law. The Copyright Act’s performers’ rights provisions appear on the pleadings Arijit Singh’s suit raised arguments about moral rights under Section 38B and the Bombay High Court noted these arguments but the courts have not resolved the AI voice cloning and deepfake question through the copyright performers’ rights framework. They have resolved it, provisionally and at the interim stage, through personality rights.
The Amitabh Bachchan Case The First Omnibus Personality Rights Order Against AI Voice Exploitation
The entry of generative AI voice cloning into Indian IP litigation was signalled, if not yet in its fullest form, by the Delhi High Court’s omnibus interim order in Amitabh Bachchan v. Rajat Nagi & Ors., CS(COMM) 819/2022, decided 25 November 2022 (2022 SCC OnLine Del 4110). Justice Navin Chawla granted an ex-parte ad-interim injunction in favour of the legendary actor, who had filed suit against nine named defendants and unnamed John Doe defendants for the unauthorised use of his name, voice, image and personality attributes to operate WhatsApp-based lottery scams and Kaun Banega Crorepati impersonation frauds and to sell merchandise and drive traffic to mobile applications and websites. The defendants were exploiting Bachchan’s deep, immediately recognisable voice one of the most distinctive in Indian cultural history alongside his photographs and the KBC brand association, to deceive the public into engaging with fraudulent schemes.
The significance of the Amitabh Bachchan ruling extends beyond its specific facts in two directions. First, it was the first time an Indian court granted what is properly described as a blanket in rem injunction for personality rights an order that restrained not just the nine named defendants but the world at large from infringing Bachchan’s personality rights through misuse of his name, voice, image and persona. This was a John Doe-style order applied to personality rights rather than copyright or trademark piracy and it fundamentally expanded the reach of the interim injunction mechanism in identity exploitation cases. Second, the court directed MeitY and the Department of Telecommunications to compel ISPs and telecom service providers to block the infringing URLs, websites and phone numbers identified by the plaintiff, integrating government regulatory authority into the enforcement of personality rights in a manner that had previously been seen primarily in trademark and copyright piracy cases. The voice, as a distinct and protectable attribute of a public figure’s identity, was for the first time placed squarely within the scope of judicially enforceable personality rights protection in India.
Anil Kapoor AI, Deepfakes and the Technology-Specific Injunction
The specificity with which Indian courts have addressed AI and deepfake technology as mechanisms of personality rights violation advanced significantly in Anil Kapoor v. Simply Life India & Ors., CS(COMM) 652/2023, decided 20 September 2023 (2023 SCC OnLine Del 6914). The case, discussed in the earlier Patenevo article on the right of publicity, is revisited here specifically for its performers’ rights and voice-related dimensions. Sixteen defendants were restrained and the injunction specifically and by name prohibited the use of artificial intelligence, machine learning algorithms, face-morphing technology and deepfake tools to create unauthorised content using Kapoor’s voice, image, likeness or persona. The Delhi High Court directed MeitY to take stringent action against the dissemination of pornographic content made using morphed images of the actor the first judicial direction of this specificity concerning AI-generated non-consensual intimate imagery in India.
For performers specifically, the Anil Kapoor judgment is important because the court grounded its intervention in the right to livelihood a dimension of Article 21 recognising explicitly that an actor’s endorsement value and commercial persona are not merely dignitary interests but economic ones and that AI-driven impersonation directly threatens the source of a performer’s commercial livelihood. This framing brings the judicial protection of performers’ identities closer in principle to the economic rights that performers’ rights legislation provides, even though the court reached this conclusion through constitutional rather than copyright doctrine.
The Landmark Case Arijit Singh v. Codible Ventures LLP
The most important and most comprehensively analysed Indian decision on AI voice cloning and performers’ rights is the Bombay High Court’s order in Arijit Singh v. Codible Ventures LLP & Ors., IPR Suit (L) No. 23443 of 2024, decided 26 July 2024 (2024 SCC OnLine Bom 2445). Justice Riyaz Chagla authored the order which has been widely described as the first Indian and among the first global judicial decisions specifically addressing the commercial exploitation of a singer’s voice and vocal style through AI-powered voice conversion technology.
Arijit Singh, at the time of filing the suit, was the most-followed artist on Spotify globally with 138.5 million monthly listeners a figure that places him above Taylor Swift and makes him the most commercially successful musical performer in the world by that metric. The thirty-eight defendants sued included AI tool developers who had built platforms specifically designed to allow users to clone Singh’s voice using retrieval-based voice conversion (RVC) technology, enabling users to generate synthetic audio recordings in Singh’s voice from any text input or source audio; bar and restaurant owners who had used such AI-cloned voice outputs as background music without authorisation; virtual reality and metaverse event organisers who had created avatar-based digital events simulating Singh’s performances; e-commerce platforms selling merchandise bearing his name and image; domain registrars who had registered arijitsingh.com and arijitsingh.in without his authorisation; and operators of platforms that allowed users to create, store, search for and share GIFs of his performances and persona without consent.
Singh’s pleadings claimed protection for a carefully enumerated set of attributes, each of which he characterised as a legally protectable facet of his personality and performers’ right. These included his voice, his vocal style and technique, his vocal arrangements and interpretations of compositions, his distinctive mannerisms in performance, his manner of singing, his physical appearance and image, his name and aliases and his signature. The pleadings engaged both the personality rights doctrine from Titan Industries and D.M. Entertainment and the performers’ rights provisions under Sections 38 and 38B of the Copyright Act, 1957, though the court’s analysis ultimately rested more comfortably on personality rights than on the copyright performers’ rights framework.
Justice Chagla held, in terms that have become the most-cited passage in Indian AI personality rights jurisprudence, that the case “shocked the conscience” of the court. The court recognised that the defendants were leveraging Singh’s global popularity and distinctive vocal identity to drive commercial traffic to their platforms and products without any authorisation or compensation. The court found that Singh’s vocal identity his specific timbre, his emotional expressiveness in performance, his instantly recognisable singing style had commercial value generated entirely by his own artistic investment over years of professional performance and that this value was being commercially extracted by the defendants without any contribution or right on their part. The court explicitly found that AI tools that allow any user to convert any audio recording into a synthesised version in Singh’s voice constitute a form of commercial exploitation of his personality that violates his right of publicity, regardless of whether any specific protected performance is reproduced. This was the court’s answer, on the facts before it, to the statutory gap identified above: the violation is not of a performers’ right in a specific performance but of a personality right in a unique and commercially valuable identity attribute the voice.
The interim injunction granted by the Bombay High Court was sweeping in both scope and media specificity. It restrained all defendants and the world at large, from using Singh’s name, voice, vocal style and techniques, mannerisms, photographs, images, signature, persona or any other aspect of his personality, for any commercial or personal purpose without his explicit consent. The order specifically covered, by name, online platforms, publications, advertisements, merchandise, domain names, generative AI tools, voice conversion technologies, synthesised voice models, digital avatars, caricatures, deepfakes, face morphing and GIFs technology categories that did not exist when the Copyright Act was last substantively amended in 2012. The injunction was made dynamic, following the template established in UTV Software v. 1337x.to for online copyright piracy, meaning it automatically extends to new manifestations of infringing conduct without requiring the plaintiff to return to court for each new variation.
Jackie Shroff The AI Chatbot Injunction and the Parody Boundary Jaikishan Kakubhai Saraf alias Jackie Shroff v. The Peppy Store & Ors., CS(COMM) 389/2024, decided 15 May 2024 (2024 SCC OnLine Del 3664) extended the Arijit Singh framework to a distinct technology category: the unlicensed AI chatbot. Among the many categories of defendants restrained by the Delhi High Court in this matter, one entity was operating an AI chatbot built specifically around Jackie Shroff’s persona, designed to simulate ongoing interactive conversation using attributes of the actor’s identity and manner of speaking. The court’s restrait of this chatbot defendant is, as noted in the broader personality rights article on this platform, the first Indian judicial order specifically prohibiting the operation of an AI conversational tool built around a celebrity’s identity without authorisation.
The Jackie Shroff judgment also drew an important line that is directly relevant to performers seeking to enforce AI-related identity rights. The court declined to restrain a YouTube channel producing content titled “Jackie Shroff is Savage, Jackie Shroff Thug Life,” characterising this as a form of comedic, tribute-oriented expressive content protected by the freedom of expression guarantee under Article 19(1)(a) of the Constitution. This refusal was a conscious application of the principle acknowledged in D.M. Entertainment and affirmed in the personality rights literature that the right of publicity, however broadly framed, does not extend to suppressing parody, satire, commentary or tribute-oriented creative expression that uses a public figure’s attributes in a non-commercial, expressive context. For performers navigating AI-related enforcement, this boundary is practically important: an injunction that sweeps up clearly expressive, non-commercial use of a performer’s identity alongside commercial AI exploitation risks Article 19(1)(a) challenge and judicial modification on that ground.
The Performers’ Rights Statutory Framework and AI The Five Critical Gaps
Drawing together the analysis of Sections 38, 38A and 38B and the AI case law, five specific gaps in the current performers’ rights statutory framework can be identified with precision. Each represents a question that the legislature must address if the Copyright Act is to function as a genuine protection for performers in the AI era rather than leaving that protection entirely to judge-made personality rights doctrine.
- The first gap is the absence of any protection for a performer’s voice as an attribute independent of a specific performance. Sections 38 and 38A protect performances. They do not protect voices. An AI voice cloning system that replicates a performer’s vocal characteristics without reproducing any specific protected performance is not addressed by the current performers’ rights provisions.
- The second gap is the absence of any right of performers to control the use of their identity in AI training data. AI voice cloning systems are trained on recordings of a performer’s voice recordings that are typically commercially released sound recordings protected by copyright under Section 13(1)(b), owned not by the performer but by the record producer. The performer’s consent to training is not currently required by either the Copyright Act or any other Indian statute and the question of whether the use of a performer’s recordings for AI training constitutes infringement of the record producer’s copyright in the sound recording or whether it attracts a fair use exception under Section 52, remains unresolved by any Indian court.
- The third gap is the absence of any right of performers to control the creation and commercial distribution of digital replicas AI-generated visual or audio simulations of their performance identity that do not reproduce any specific protected performance. Deepfakes and AI-generated visual avatars are not addressed anywhere in the Copyright Act, 1957.
- The fourth gap is the term limitation in Section 38, which provides that performers’ rights subsist for fifty years from the year in which the performance was made. AI voice cloning systems can plausibly clone a performer’s voice from recordings made decades ago. A performance made in 1980 would have had its performers’ rights expire by 2031 at the latest meaning that AI exploitation of recordings from the earlier decades of a performer’s career may fall entirely outside the temporal scope of Section 38 protection.
- The fifth gap is the absence of any collective management infrastructure specifically designed for performers’ AI-related rights. The Copyright Act’s provisions on copyright societies under Section 33 permit the formation of collective licensing bodies and the Indian Singers’ Rights Association (ISRA) has been judicially recognised as an appropriate body for collecting royalties for commercial use of performers’ recorded vocal performances. But ISRA’s current mandate and licensing framework does not extend to AI voice cloning or synthetic performance exploitation there is no licensing mechanism through which an AI platform could legitimately license the use of a performer’s vocal attributes for voice model training or for the commercial distribution of AI-generated content and therefore no royalty stream that performers can currently access from this form of use.
The WPPT Standard and India’s International Obligations
India ratified the WIPO Copyright Treaty (WCT) and the WIPO Performances and Phonograms Treaty (WPPT) in 2018 and the Copyright (Amendment) Act, 2012, was in substantial part directed at bringing India’s domestic performers’ rights law into compliance with these treaties. Article 5 of the WPPT grants performers moral rights in their live aural performances and performances fixed in phonograms and Article 6 grants exclusive economic rights of broadcasting and communication to the public for unfixed performances. Articles 7 through 10 grant performers exclusive reproduction, distribution, rental and making-available rights in their phonogram performances. India’s Section 38A implements these WPPT provisions and Section 38B implements Article 5’s moral rights. However, the WPPT itself was negotiated in 1996 and predates generative AI by several decades. It does not address AI voice cloning, digital replicas or the training of AI systems on performers’ recordings. The question of whether the WPPT’s right of “making available” under Article 10 which the WIPO Standing Committee on Copyright and Related Rights has more recently discussed in the context of AI-generated content extends to the making available of AI-generated voice clones of performers is an open one at both the international and domestic level.
The Information Technology Act, 2000 and the Intermediary Dimension
Beyond copyright, the Information Technology Act, 2000 and specifically Section 79 and the Intermediary Guidelines and Digital Media Ethics Code, 2021, have become relevant to AI voice cloning and deepfake enforcement in two distinct ways. First, platforms that host AI voice conversion tools or deepfake-generated content can potentially claim the safe harbour protection of Section 79, which protects intermediaries from liability for third-party content where the intermediary has not initiated the infringing conduct and complies with takedown requests on actual knowledge. The application of Section 79 to AI platforms that actively provide voice cloning tools as opposed to passively hosting user-generated content mirrors the question the Delhi High Court addressed in the DRS Logistics Google AdWords litigation: an AI platform that designs and provides a tool specifically for generating voice clones of identifiable individuals may be too actively involved in the infringing conduct to credibly claim passive intermediary status. Second, under the Intermediary Guidelines, 2021, platforms are required to implement complaint mechanisms and to remove content reported as non-consensual, including fake or morphed content involving individuals without their consent. The Arijit Singh and Anil Kapoor cases both involved directions to MeitY to compel ISPs to implement blocking orders a pattern that demonstrates the convergence of IT Act regulatory authority and IP court jurisdiction in AI deepfake enforcement.
Comparative International Context The US RIGHT Act and the EU AI Act
The international legislative response to AI voice cloning and deepfakes as performers’ rights issues is developing, though no comprehensive binding international framework yet exists. In the United States, the No Fakes Act of 2024, which remains under legislative discussion, specifically proposes a federal right for individuals including performers to control the creation and commercial distribution of AI-generated digital replicas of their voices and visual likenesses. Several US states, including Tennessee through the Ensuring Likeness Voice and Image Security (ELVIS) Act, 2024, have enacted state-level performers’ voice cloning protections. The European Union’s AI Act, 2024, classifies deepfake generation systems as high-risk AI applications requiring specific disclosure and labelling obligations. None of these frameworks has an exact Indian equivalent and each reflects the specific legal and constitutional traditions of the jurisdiction that enacted it. But their collective existence demonstrates a global legislative consensus, still forming, that performers’ rights in the AI era require specific statutory intervention that existing copyright and personality rights frameworks cannot adequately provide.
The Reform Agenda What Indian Law Needs
The deficiency in India’s current legal framework for protecting performers against AI voice cloning and deepfakes is not a failure of judicial creativity the courts have been remarkably responsive, have developed workable interim relief mechanisms and have applied available doctrines flexibly and intelligently to new technological contexts. The deficiency is legislative. The Copyright Act, 1957, needs specific amendment to address four concrete requirements. First, the definition of “performer’s right” in Section 38 should be extended to protect a performer’s voice and distinctive vocal characteristics as an attribute independent of any specific performance, alongside rights over their visual likeness and identity attributes generally. Second, a specific prohibition on the creation and commercial distribution of AI-generated digital replicas voice clones, visual deepfakes, synthetic performances without the performer’s consent should be inserted, with appropriate exceptions for parody, satire and genuine artistic commentary. Third, the use of performers’ recordings as training data for AI systems should require the consent of both the record producer as copyright holder and the performer as rights holder in the performance, with an appropriate royalty or benefit-sharing mechanism. Fourth, Section 33 of the Act should be amended to enable existing copyright societies, particularly ISRA, to extend their collective licensing mandates specifically to AI-related uses, creating a workable licensing framework that can provide legitimate commercial access to performers’ attributes for AI development in exchange for fair compensation.
Conclusion
India’s performers’ rights framework under Sections 38, 38A and 38B of the Copyright Act, 1957, was a significant legislative achievement when it was substantially reformed in 2012. It gave Indian performers, for the first time in the country’s copyright history, genuine economic and moral rights in their live performances and it aligned Indian law with the WPPT standards that govern performers’ rights internationally. It has, however, been left behind by a technology that the 2012 amendments could not have anticipated. The AI voice cloning that made Arijit Singh’s case necessary, the deepfake technology that drove Anil Kapoor’s 2023 injunction, the AI chatbot that prompted Jackie Shroff’s 2024 proceedings none of these can be cleanly addressed within the current performers’ rights statutory framework. The courts have responded by creative deployment of personality rights doctrine, constitutional privacy jurisprudence and John Doe injunction mechanics. That response has been effective at the interim stage and has produced a body of case law that, collectively, amounts to a powerful judicial statement that AI exploitation of performers’ identities without consent is legally impermissible in India. But judge-made doctrine built on provisional interim orders is not a substitute for legislative clarity, particularly for the millions of performers who are not famous enough to sue a technology company in the Bombay High Court and obtain a dynamic omnibus injunction from Justice Riyaz Chagla. The reform of India’s performers’ rights statute for the AI era is, accordingly, one of the most urgent items on the Indian IP legislative agenda.
References
- Copyright Act, 1957 (as amended by Copyright Amendment Act, 2012) Sections 38, 38A, 38B – https://copyright.gov.in/Documents/CopyrightRules1958.pdf
- Arijit Singh v. Codible Ventures LLP & Ors., IPR Suit (L) No. 23443 of 2024, Bombay High Court, 26 July 2024 (2024 SCC OnLine Bom 2445) – https://indiankanoon.org/doc/168421456/
- Amitabh Bachchan v. Rajat Nagi & Ors., CS(COMM) 819/2022, Delhi High Court, 25 November 2022 (2022 SCC OnLine Del 4110) – https://indiankanoon.org/doc/23523817/
- Anil Kapoor v. Simply Life India & Ors., CS(COMM) 652/2023, Delhi High Court, 20 September 2023 (2023 SCC OnLine Del 6914)- https://indiankanoon.org/doc/113724486/
- Jaikishan Kakubhai Saraf v. The Peppy Store & Ors., CS(COMM) 389/2024, Delhi High Court, 15 May 2024 (2024 SCC OnLine Del 3664) – https://indiankanoon.org/doc/165756699/
- Fortune Films International v. Dev Anand, AIR 1979 Bom 17 – https://indiankanoon.org/doc/1636640/
- WIPO Performances and Phonograms Treaty (WPPT), 1996 – https://www.wipo.int/treaties/en/ip/wppt/
- Information Technology Act, 2000 Section 79 – https://legislative.gov.in/sites/default/files/A2000-21.pdf
- Intermediary Guidelines and Digital Media Ethics Code, 2021 –https://legislative.gov.in/sites/default/files/GSR139E.pdf
- WIPO AI Voice Cloning and Personality Rights (Magazine Article on Arijit Singh) – https://www.wipo.int/en/web/wipo-magazine/articles/ai-voice-cloning-how-a-bollywood-veteran-set-a-legal-precedent-73631
- National IPR Policy, 2016 DPIIT – https://dpiit.gov.in/sites/default/files/National_IPR_Policy_English.pdf
- Copyright Office India Official Portal – https://copyright.gov.in
Frequently Asked Questions:
Q1. Does Indian copyright law protect performers against AI voice cloning?
The Copyright Act 1957 protects performers’ rights in specific live performances under Sections 38, 38A and 38B. However these provisions do not protect a performer’s voice as a general attribute independent of a specific performance. AI voice cloning that replicates vocal characteristics without reproducing a specific protected performance falls outside the current performers’ rights statutory framework.
Q2. What was decided in Arijit Singh v. Codible Ventures LLP 2024?
The Bombay High Court granted a sweeping omnibus injunction restraining all defendants and the world at large from using Arijit Singh’s name, voice, vocal style, mannerisms, persona or any other attribute of his identity for commercial or personal purposes without his explicit consent. The court specifically prohibited AI voice conversion tools, deepfakes, digital avatars and GIFs — the first Indian judicial order of this specificity targeting AI voice cloning technology.
Q3. What are performers’ rights under Section 38A of the Copyright Act 1957?
Section 38A grants performers exclusive economic rights to make sound recordings or visual recordings of their performance, reproduce such recordings, broadcast the performance and sell or hire copies of the recordings. These rights were substantially enhanced by the Copyright Amendment Act 2012 in compliance with India’s obligations under the WIPO Performances and Phonograms Treaty.
Q4. Can an AI-generated deepfake of a performer violate Indian law?
Yes. Even though the Copyright Act does not specifically address deepfakes, Indian courts have granted injunctions against deepfake creation under the constitutional right to privacy under Article 21, the common law tort of passing off and the judge-made right of publicity. The Delhi High Court in Anil Kapoor v. Simply Life India 2023 specifically prohibited the use of AI and deepfake tools to create unauthorised content using a performer’s identity.
Q5. What legal reforms does India need to protect performers from AI exploitation?
India needs to amend the Copyright Act 1957 to protect a performer’s voice and visual identity as attributes independent of specific performances, prohibit AI-generated digital replicas without performer consent, require performer consent for AI training data use, and empower copyright societies like ISRA to create licensing frameworks specifically for AI-related uses of performers’ attributes.
Compulsory Licensing Copyright Act Copyright Act 1957 copyright case laws Copyright Enforcement Copyright Infringement Copyright Law Copyright Rules Deceptive Similarity Descriptive Marks India Indian IP Framework Indian Patent Law Indian Trademark Law Intellectual Property Law IP Law India Patent Claims Patent Enforcement Patent Infringement Patent law Patent Revocation Patents Act Patents Act 1970 Section 9 Trade Marks Act Section 29 The Patent Act 1970 Trademark Infringement Trademark Registration Trade Marks Act 1999 Trade Marks Rules 2017 TRIPS Compliance