Generative AI can now reproduce a person's face, voice and mannerisms with a fidelity that the average viewer cannot tell apart from reality. What began as novelty content—face-swapped videos, synthetic voice clips circulated for entertainment—has, since 2023, become a live commercial and reputational threat. Bollywood actors and playback singers have gone to court in rapid succession over AI-generated impersonations used in merchandise, advertising, chatbots and, in some cases, pornographic content. India has no standalone “right of publicity” statute to answer this. What it has instead is a patchwork of overlapping legal regimes—three civil doctrines, with criminal and copyright law layered over them—never designed to work together, that courts have stitched into a functioning if fragile protection framework.
Three Borrowed Doctrines, One New Problem
The first and most heavily used route is the common law tort of passing off, extended by Indian courts from protecting a trader's goodwill in goods to protecting a celebrity's goodwill in their own persona. The logic: an unauthorised use that creates a false impression of endorsement damages a celebrity's reputation exactly as a counterfeit product damages a manufacturer. Running alongside it is the constitutional right to privacy and dignity under Article 21, which the Supreme Court recognised in Justice K.S. Puttaswamy v. Union of India, (2017) 10 SCC 1, building on R. Rajagopal v. State of Tamil Nadu, (1994) 6 SCC 632, giving individuals control over their own identity and image even outside a purely commercial context. Because fundamental rights are enforceable primarily against the State, courts deploy privacy and dignity against private defendants as the normative foundation for injunctive relief rather than as a free-standing cause of action. The third route is statutory: trade mark law. Section 14 of the Trade Marks Act, 1999 permits the Registrar to require a living person's consent before registering a mark that falsely suggests a connection with them—a bar on registration rather than a right of action. Separately, the Act's broad definition of “mark” lets a celebrity register their name or signature as a trademark in its own right—and, once registered, bring infringement remedies under Section 29 alongside the common law passing-off action Section 27(2) preserves.
Sitting on top of all three is the criminal law. Section 66C of the Information Technology Act, 2000 (IT Act) penalises identity theft, and Section 66D cheating by personation using a computer resource—squarely applicable to a deepfake or AI chatbot built to impersonate a real person—while Sections 67 and 67A address obscene or sexually explicit material, relevant wherever the deepfake is pornographic. The Bharatiya Nyaya Sanhita, 2023 adds parallel offences for cheating, forgery and defamation. A single deepfake of a public figure can therefore trigger a civil passing-off claim, a privacy and dignity claim, a trade mark infringement claim, a copyright claim (discussed below), and criminal liability, all at once—regimes with different proof standards, different remedies, and none of them drafted with AI in mind.
The Cases That Built the Doctrine
The foundation was laid well before deepfakes existed. ICC Development (International) Ltd. v. Arvee Enterprises, 2003 (26) PTC 245 (Del) established that publicity rights vest in an identifiable individual, not an event or organisation—rejecting the ICC's own claim over the 2003 Cricket World Cup on that basis. Titan Industries Ltd. v. Ramkumar Jewellers, 2012 (50) PTC 486 (Del), linked that principle to passing off: a jeweller who copied Titan's advertisement featuring Amitabh and Jaya Bachchan was liable even without a registered trade mark in the Bachchans' image, because consumers would assume an endorsement that was never given. D.M. Entertainment Pvt. Ltd. v. Baby Gift House, CS(OS) 893/2002 (Delhi High Court, April 29, 2010), decided two years before Titan, had already gone further, protecting singer Daler Mehndi's likeness reproduced without authorisation on dolls as a quasi-property right, and holding that even non-competing goods can infringe where they trade on a celebrity's established goodwill.
That doctrinal base is what Indian courts reached for when AI-generated impersonation arrived. In Amitabh Bachchan v. Rajat Nagi & Ors., CS(COMM) 819/2022 (Delhi High Court, November 25, 2022), the court granted a John Doe injunction protecting the actor's name, voice and image. Ten months later, in Anil Kapoor v. Simply Life India & Ors., CS(COMM) 652/2023 (Delhi High Court, September 20, 2023), Justice Prathiba M. Singh granted an ex parte injunction after finding the actor's name, voice, image and catchphrase were being exploited through AI face-morphing, GIFs and pornographic content—ordering domain suspension and the blocking of 44 links. It was the first major Indian ruling to confront AI-enabled misuse head-on. Eight months later, Jaikishan Kakubhai Saraf alias Jackie Shroff v. The Peppy Store & Ors., CS(COMM) 389/2024 (Delhi High Court, May 15, 2024) extended protection to an unlicensed AI chatbot built on the actor's persona while pointedly declining to restrain genuine meme and parody content, recognising its independent creative value. And in Arijit Singh v. Codible Ventures LLP, COM IPR Suit (L) No. 23443 of 2024 (Bombay High Court, July 26, 2024), the court delivered one of India's first rulings directed squarely at AI voice cloning, holding that offering a tool that converts any voice into a celebrity's violates the right to control one's own voice, and granting an ad-interim injunction that, in the court's words, “may also operate as a dynamic injunction”, drawing on the “dynamic” injunctions first developed in the Delhi High Court's piracy jurisprudence (UTV Software Communications Ltd. v. 1337x.to, CS(COMM) 724/2017, Delhi High Court, April 10, 2019), under which orders can be extended to newly identified infringing sites on application rather than through a fresh suit.
Since then, the trickle has become a flood. In 2025 the Delhi High Court extended similar protection to Sadhguru (June); to Karan Johar, Aishwarya Rai Bachchan (Aishwarya Rai Bachchan v. Aishwaryaworld.com, CS(COMM) 956/2025) and Abhishek Bachchan (September), the Bachchans' pleadings citing AI chatbots, face-swapped imagery and YouTube channels running deepfake videos; and to Nagarjuna, while the Bombay High Court protected Asha Bhosle (October 2025). Interim orders for Hrithik Roshan, Ajay Devgn and others have followed into 2026.
Read together, the trajectory is unmistakable: from recognising publicity rights in principle, through applying passing off to unauthorised merchandise and endorsement, to explicitly confronting AI-generated likeness across image, voice and conversational agents each decision extending the same underlying logic to a new technological modality, well ahead of any legislative definition.
The Copyright Angle
Copyright law offers a further, if narrower, line of defence. The Copyright Act, 1957 does not protect a face or a voice as such—neither is a “work”—but it does protect the material deepfakes are built from. Most synthetic clips are produced by reproducing and altering existing films, photographs and sound recordings, which can infringe the owner’s exclusive rights under Section 14 and constitute infringement under Section 51, unless the use falls within the narrow fair-dealing exceptions in Section 52. More directly relevant to the individual are performers’ rights: Section 38A gives performers exclusive control over the recording and communication of their performances, and Section 38B grants them moral rights to be identified and to restrain any distortion or modification of a performance that would prejudice their reputation—a natural fit for a cloned voice or a manipulated on-screen appearance.
The limits are equally clear. Copyright in a film or song usually vests in the producer or label, not the person depicted, so the celebrity may not even own the right being infringed; performers’ rights attach to an actual performance, not to newly generated content that merely imitates a voice or style; and whether training AI models on copyrighted material is itself infringement remains unsettled in India, a question now squarely raised before the Delhi High Court in ANI Media Pvt. Ltd. v. OpenAI, and echoed in the Bachchans' 2025 suits seeking an order that Google prevent deepfake videos uploaded to YouTube from being used to train other AI models. Copyright is therefore best used as a supporting claim alongside passing off and personality rights, rather than a substitute for them.
Where the Gaps Remain
The improvisation has real limits. Virtually every reported case so far involves a public figure with the resources to obtain urgent High Court relief; an ordinary individual whose likeness is misused is left largely with the slower, less tailored criminal-law route, and with the still-untested possibility of treating a person's face and voice as personal data under the Digital Personal Data Protection Act, 2023. There is also no codified standard for what makes a persona “protectable” or how distinctive it must be—the AI-era rulings have almost all been interim, often ex parte, orders under the court's general injunctive power, not considered judgments under a dedicated statute, so the underlying legal test stays undefined and inconsistent across benches. That may soon change: since September 30, 2026, the Delhi High Court has
been hearing three personality rights suits together—filed by Vivek Oberoi, Janhvi Kapoor and Aaradhya Bachchan—to settle what the right covers, including whether a claimant must show that someone profited from their persona. And because deepfake tools and hosting infrastructure sit offshore, even the dynamic injunction model, now routinely sought in persona cases, is itself an admission that a one-time order cannot keep pace with mirror sites—the same cross-border enforcement problem that has dogged Indian internet law for years.
The 2026 IT Rules: Regulation Catches Up
The regulatory response, foreshadowed by government advisories on deepfakes in 2023 and 2024, took binding form only this year. Effective February 20, 2026, the Ministry of Electronics and Information Technology's amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 introduced a defined category of “synthetically generated information”—content that is artificially or algorithmically created or altered to appear authentic—which must now carry clear, prominent labelling, a prominently prefixed audio disclosure for audio content, and embedded metadata that platforms cannot allow users to strip. Significant social media intermediaries must obtain user declarations on whether uploaded content is synthetic and verify them through technical measures; the deadline for acting on a court order or government notice drops from 36 hours to three, and to just two hours for complaints of non-consensual intimate imagery, including morphed and deepfake content; and non-compliance risks the loss of Section 79 safe-harbour protection. It is the first time Indian law has expressly defined and regulated AI-generated synthetic content—converting what courts had been doing informally through injunctions into a standing regulatory obligation.
The Road Ahead
What the 2026 Rules do not do is define the underlying right. They regulate the distribution of synthetic content, not the question of who holds a personality right, what it protects, and for how long—a gap that a dedicated statute (whether standalone, or a defined chapter within the Trade Marks Act or the long-proposed but still unintroduced Digital India Act) is better placed to close, along with codifying the parody and free-expression carve-out that Jackie Shroff created only informally, and widening access beyond litigants who can afford a High Court suit. Until that happens, protection for any individual will keep depending on how quickly, and in which High Court, an injunction can be secured—a workable outcome so far, but a fragile foundation for a right that generative AI has made more valuable, and more vulnerable, than ever.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
