Copyright Law And Deepfakes - Need For A Sui- Generis Law

Author: LawSphereX

Category: Intellectual Property

Published: 2026-08-23T18:58:55.719Z

Updated: 2026-08-30T07:12:21.095Z

COPYRIGHT LAW AND DEEPFAKES: NEED FOR A SUI-

GENERIS LAW?

Poorva Singhal, National Law University, Delhi

ABSTRACT

This article examines whether India’s existing Copyright Act, 1957, and ancillary remedies can meaningfully address the harms posed by photorealistic visual deepfakes, or whether a narrowly tailored sui generis personality right is required. Motivated by the recent spate of celebrity litigation and the use of deepfakes in election-period misinformation, the paper uses doctrinal analysis of statutory text and case law (including the Sehgal–Rewal split), alongside comparative study of Denmark’s recent amendment, to map the gaps in India’s current regime.

The paper shows that moral-rights protection under Section 57, performer protections under Section 38, privacy and passing-off torts, and IT Act provisions are each useful but fragmentary: they fail to capture wholly synthetic imitations, hinge on fact-sensitive showings (reputation, commercial confusion, authorship), and suffer from doctrinal uncertainty and slow, costly enforcement. Denmark’s IP-style approach offers clearer substantive coverage, streamlined takedown and damages remedies, and broader access to relief, but risks commodifying personhood if transplanted wholesale.

The paper therefore proposes a calibrated Indian design: a statutory definition limited to non-consensual, highly realistic imitations; universal standing for takedown and compensation; strengthened intermediary duties; and built-in safeguards for satire, journalism, free expression, and the inalienability of core personality rights. The reform would reduce litigation costs, improve enforceability, and better protect dignity without unduly restricting speech.

Introduction

The digital age narrows the boundary between what is public and how it can be manipulated or its likeness exploited. India is recently experiencing an escalating proliferation of hyperrealistic “deepfakes” which are harming the country’s information integrity and public discourse. This was severely used during the elections to manipulate voters and spread misinformation. 1 Unauthorised copies of artists desecrate the very sanctity of the artist and his/her work and destroy the public trust. This problem grows especially pronounced when there is no dedicated “personality rights statute” to protect individuals.

The recent string of cases filed by famous celebrities such as Aishwarya Rai Bachchan, Abhishek Bachchan, Karan Johar, Asha Bhosle and Kumar Sanu against content creators and platforms that host unauthorised deepfake content reveals the gap in law in the current age of artificial intelligence and what it means to protect one’s identity in digital spaces. This paper examines whether the existing provisions within the Copyright Act, 1957, can combat the problem of visual deepfakes or whether a sui-generis law is needed. Through existing provisions and case laws, I will argue that the existing provisions provide very limited protection against non-consensual visual deepfakes and a targeted sui generis right drawn from Denmark’s IP-style model, but calibrated for free speech would be a more holistic solution. Indian Copyright Law S.

57 of the Copyright Act provides moral rights to the author independent of the author’s economic rights. 2 The importance of moral rights was emphasised in Amarnath Sehgal v. Union of India, where the Delhi HC stated that moral rights are “considered the ‘soul’ of the author’s creation”. 3 This protects authors if a copyrighted performance is manipulated without permission. However, in the context of visual deepfakes, this protection is very narrow, as it only protects existing copyrighted material and not standalone attributes, such as a person’s voice timbre or facial features. 4 This is because Indian copyright law does not contain any 1 Jaskirat Singh Bawa, “From Misinformation To Fraud, The Alarming Rise Of Deepfakes In India” NDTV (17 July 2024)

2 Copyright Act 1957 (India) s 57.

3 2005 (30) PTC 253 (Del)

4 Krunal Mehta, “The Persona Paradox

Personality Rights, Deepfakes & Identity in Indian Law” King Stubb & Kasiva (6 October 2025) explicit protection for “personality rights”. However, the Delhi High Court has stated that the right of publicity is exclusive to a person or any manifestations of his/her persona, such as their name, personal attributes, autograph, or vocal characteristics. Consequently, the individual concerned has the sole right to derive pecuniary benefits from the right of publicity. 5 S. 38 provides additional protection for the right of performers against unauthorised fixation or broadcasting of performances, such as a singer’s live show. 6 However, it does not protect the identity or the persona of the performer, which limits its value against AI impersonators. Thus, the statutory scheme shows limits when the deepfake is wholly synthetic, i. e. completely generated by an AI model.

Copyright’s originality/authorship tests presuppose a human author and a protectable work. Due to this, an AI-generated photorealistic likeness of a living person typically does not qualify as a “work” authored by the person whose likeness appears, nor does it necessarily reproduce a prior copyrighted fixation. This creates a structural “authorship/source” gap in the Act’s coverage. 7 Another problem arises in the opposing interpretation of what counts as “mutilation” deserving of protection under the Copyright Act by the Delhi High Court in Amarnath Sehgal and Raj Rewal. 8 The former adopts a broad purposive reading of S. 57, holding that “mutilation” could include outright destruction or any such destruction of the work which could be prejudicial to the author’s reputation.

9 On the other hand, Raj Rewal qualified moral rights with public or property interests, holding that a property holder has the right to dispose of their property and moral rights cannot override constitutional property rights. 10 This differentiation in interpretation creates a hurdle for getting copyright protection, even in situations where deepfakes modify a pre-existing copyrighted work. In situations where the platform lawfully owns the media content, Rehwal is less likely to provide relief. Common Law and Other Remedies India relies on a patchwork of common law and constitutional law protections, including Article 21, privacy jurisprudence highlighted by Puttaswamy v. Union of India, and tort law

5 ICC Development Ltd. v. Arvee Enterprises & Anr, 2003 (26) PTC 245 Del

6 Copyright Act 1957 (India) s 38

7 Praditya Singh Narolia and Ramya Manoj Panichker, “Deepfakes in India

A Legal Labyrinth of Copyright and IPR” JOUR (27 April 2025)

8 Raj Rewal v Union of India & Ors (Delhi High Court, 28 May 2019) CS (COMM) No. 3/2018

9 Amarnath Sehgal, n 3

10 Raj Rewal, n 7

doctrines such as passing off. 11 As per Puttaswamy, an individual’s right to privacy encompasses their ability to exercise control over the dissemination of personal information, which is both necessary and imperative. This control is compromised by the usage of deepfakes as it encroaches upon an individual’s personal autonomy. 12 Certain provisions of the Information Technology Act, 2000 also protect against violation of privacy, explicit content and impersonation using a computer. 13 Passing off is a common law tort used to enforce unregistered trademark rights. Traditional cases involving this remedy involve situations where one party misrepresents their goods or services as those of another, limiting the damage to potential financial loss. In deepfakes, the scope of damage is expanded to public image and personal brand.

14 Additionally, in the past, passing off required a certain level of effort and intent. Deepfakes, by contrast, can be created and disseminated with alarming ease, often by individuals without significant resources or traditional commercial motives. 15 This also limits the remedy of passing off, which requires a link to commercial activity or confusion. Unjust Enrichment occurs when someone profits from using another person’s persona without permission at the plaintiff’s expense. Practically, the courts issue broad injunctions and “John Doe” actions to preempt misuse. 16 India’s solution to deepfakes, which is illustrated by a string of Delhi High Court cases regarding personality rights filed by celebrities, is to craft hybrid remedies drawing on these provisions. This results in uneven protection: while present and can be granted, but fragile in the absence of any explicit statute.

11 Ruhi Khan, “From deepfakes to dignity: what Bollywood’s personality rights battle with AI tells us” (14 October 2025) LSE Blog 12 Shinu Vig, “Regulating Deepfakes: An Indian Perspective” Journal of Strategic Security 17, no. 3 (2024): 70 -93

13 Information Technology Act 2000 (India), s 66E, s 66D, s 67

14 Vishnu Vijandran, “The Intersection of Deepfakes and Intellectual Property Law

Applying the Concept of Passing Off” (11 March 2024) Aqran Vijandran Advocates & Solicitors 15 Lutz Finger, “Overview Of How To Create Deepfakes - It’s Scarily Simple” (8 September 2022) Forbes

16 Yagay AndSun, “Deep-Fake

Copyright on Face and Voice - International Laws. WIPO. Indian IPR Organisation” (4 October 2025) TaxTMI Danish Amendment for Protection Against Deepfakes The Danish government changed the copyright law to protect every individual against deepfakes, defining the term as “a very realistic digital representation of a person, including their appearance and voice. ” The Danish Law gives persons a clear legal basis to demand takedown of unauthorised deepfakes, seek compensation and, in targeted cases, hold hosting platforms to account. 17 It creates an exception for imitations that are mainly expressions of caricature, satire, parody, pastiche, criticism of power, social criticism, etc. , unless the imitation constitutes misinformation that may cause specific and serious danger to the rights or essential interests of others.

18 This approach improves upon the Indian approach of providing remedies through a scattered bunch of remedies. Firstly, it provides a substantive clarity by creating a single statutory cause of action tied to the person’s likeness itself. Where Indian claimants currently must identify the “best” fit for their problems, Danish law simplifies liability by making the protection intrinsic to the individual’s persona. That clarity reduces doctrinal uncertainty and the transaction costs of litigation. 19 Secondly, it provides procedural efficiency and enforceability. The Danish law provides takedown mechanisms and enforcement pathways wherein people affected by deepfake content can request its removal and can demand compensation for unauthorised use of their likeness. This right would extend to 50 years beyond the individual’s death.

20 Additionally, if the platform does not remove illegal content after receiving a notification, there may be financial consequences for the provider under the EU Digital Services Act. Rather than piecing together John-Doe orders and cross-border disclosure requests, victims obtain a recognised IP remedy that can trigger prompt platform action and statutory remedies for non-compliance. Thirdly, the Denmark law provides a universal coverage unlike the piecemeal Indian remedy, which only provides remedies to celebrities. This equality of remedy reduces the “access to 17 Miranda Bryant, “Denmark to tackle deepfakes by giving people copyright to their own features” (27 June 2025) The Guardian

18 Martin Dahl Pedersen and Anne-Sophie Kofoed Rasmussen, “Fighting deepfakes through the Danish

Copyright Act” (8 July 2025) Kromann Reumert

19 Lov Om Ophavsret (Consolidated Act No. 1093 of 20 August 2023), s 73a

20 Andrea Willige, “Deepfake legislation

Denmark moves to protect digital identity” (30 July 2025) World Economic Forum justice” gap that leaves private persons disproportionately vulnerable. Coupled with platform obligations and potential fines for non-compliance, the Danish model strengthens deterrence against creators and hosts of harmful deepfakes. However, it is important to note that the Danish law is limited to Denmark, and the illegal content may therefore still be accessible from other countries. 21 However, critics of the Danish Act state that keeping “copyright” as the core of the law would commodify identity rather than simply making it a focus of protection. This can be seen in how the act conceives the new right to control deepfakes as a marketable, exploitable right, subject to monetisation by way of licensing. This conveys the message that deepfakes are not taboo but rather a licensing opportunity.

There is an additional critique that the focus should be on the enforcement of the already existing remedies rather than the creation of new ones to eradicate deepfakes. 22 Sui Generis Approach? The existing Indian remedies are insufficient to meet the scale and specificity of harm that deepfakes create for reasons given above. S. 57 of the Copyright Act protects the integrity of a fixed copyrighted work and can curb derogatory edits to an author’s recorded performance. Additionally, constitutional law and tort doctrines provide important complementary routes. But these are fragmentary: they require either a pre-existing protected “work” (making wholly synthetic deepfakes fall outside s.

57), or proof of commercial confusion, reputational prejudice, or a serious privacy invasion, which are fact-sensitive, resource-intensive and often slow to deliver takedowns against anonymous or cross-border hosts. Section 57’s scope is further muddied by divergent case law (Sehgal’s expansive “mutilation” reading versus Rewal’s contextual limits) producing doctrinal uncertainty for claimants. Denmark’s statutory intervention addresses these shortcomings by creating a single, personality-centred cause of action: realistic AI imitations of a person’s face, voice or body are actionable, victims can demand takedown and compensation, and platforms face enforceable duties to remove illicit content (with defined carve-outs for parody/journalism). By embedding protection in the copyright framework, Denmark streamlines enforcement along familiar

21 Fighting Deepfakes through the Danish Copyright Act, n 18

22 P. Bernt Hugenholtz, “Deepfake Bills in Denmark and the Netherlands

Right idea, wrong legal framework” (6 August 2025) Kluwer Copyright Blog notice-and-remove lines and reduces the need for patchwork litigation strategies. That said, a Danish-style sui generis right should not be transposed blindly to India. A workable Indian design would combine three core features: 1) a statutory definition of “visual deepfake” limited to highly realistic, non-consensual imitations likely to cause reputational, financial or safety harms (with an explicit exclusion for bona fide satire, academic research and journalism) Alternatively, a standalone Personality Rights Act could codify what traits are protected and under what conditions. Such a law should specify covered attributes (name, face, voice, catchphrases, style) and define when unauthorised use is actionable (e. g.

for commercial exploitation or reputational harm); 2) The reform should allow anyone (not just celebrities) to demand removal of deepfakes of themselves, regardless of the content’s intent. This would acknowledge that even non-malicious deepfake use (e. g. forgery or identity spoofing) undermines personal dignity23; and 3) extend intermediary liability rules (such as Section 79 of the IT Act) to require quick action against flagged deepfakes. 24 Drawing from the EU Digital Services Act, India could mandate that platforms use automated filters or tagging systems to detect AI-generated content and impose penalties for non-compliance. 25 In parallel, there could also be a strengthening of already existing moral rights by clarifying that unauthorized AI-imitation of a recorded performance would be treated as a prejudicial distortion under Section 57. S.

38 could be expanded so that even AI “covers” of a performer’s voice or image infringe the original artist’s rights. This would give creators of original work (songs, films) more leverage to block unauthorized AI clones that copy them. Additionally, this would also clarify explicitly about the extent of copyright protection over existing copyrighted material. Any sui generis right would need to be accompanied by key safeguards. Firstly, protect free expression by narrowly drafting the exceptions for political speech, parody, and public-interest journalism, and by requiring claimants to show significant harm for injunctions. Secondly, avoid commodifying personhood by making core personality rights inalienable even while allowing limited licensing for consensual uses.

Thirdly, align the law with data-protection rules 23 Anonymous, “Denmark Gives everybody the right to their own body, facial features and voice to counter Deepfakes” (23 August 2025) The Good Lobby

24 Information Technology Act 2000 (India), s 79

25 Tambiama Madiega, “Generative AI and watermarking” (December 2023) European Parliamentary Research Service PE 757. 583 so remedies coincide rather than conflict. Finally, built-in judicial review and periodic legislative review clauses to calibrate the regime as technology evolves. Conclusion The law we have today is a patchwork which useful in places, dangerously thin in others. S 57 and related provisions can curb an edited clip or a doctored performance, privacy and passingoff doctrines can blunt some harms, and injunctive John-Doe orders work in high-profile instances. But when a photorealistic video is born in an algorithm rather than from a single “work,” these tools misfire: the authorship logic of copyright fails, privacy claims are factsensitive and slow, and passing-off demands a commercial hook that many malicious deepfakes simply lack.

The Sehgal–Rewal split over what counts as “mutilation” neatly illustrates the deeper problem that judges can stretch moral rights to protect reputation in some contexts and then pull back when property or public interests loom. That unpredictability is exactly what deepfake creators can exploit. This illustrates the need of the hour for India, needing more than better litigation tactics. A narrowly tailored sui generis right would achieve three things in one stroke. Firstly, it would provide victims with a clear, headline cause of action when highly realistic, non-consensual visual imitations threaten reputation, safety or financial interests. Secondly, it would channel enforcement through familiar IP-style takedown and damages mechanics so platforms respond quickly. Thirdly, it would democratise access to relief so private persons aren’t left to the mercy of costly lawsuits.

However, any statutory fix must be careful to ensure that the right to free speech is not undermined and that personality rights are not reduced to a marketable commodity.