If 2026 has a defining IP story in India, it isn’t a patent dispute or a trademark battle. Instead, it’s the sheer volume of public figures who have gone to court this year to stop AI from impersonating them, and what that wave of litigation reveals about how personality rights in India actually work in practice. Bollywood actors, cricketers, a religious leader, a business tycoon, a playback singer, a journalist and a politician have all sought — and won — interim protection from Indian courts against AI-generated deepfakes, voice clones and unauthorised commercial exploitation of their identity. The pattern is now too large and too consistent to treat as a string of one-off celebrity disputes. It is a legal trend with direct consequences for how any brand, agency or public-facing individual should be thinking about identity, AI, and risk.
The Scale of the 2026 Wave of Personality Rights
The roll call is genuinely wide-ranging. Actor Preity Zinta approached the Bombay High Court, which ordered Meta, Google LLC, X Corp and several websites to remove AI-generated deepfake videos, morphed photographs and an AI chatbot persona built on her likeness. Actors Aishwarya Rai Bachchan, Abhishek Bachchan and filmmaker Karan Johar separately petitioned the Delhi High Court, which barred the creation, sharing or sale of digital content and merchandise using their names, images, voices or personas — including, in Johar’s case, protection for nicknames and memes associated with him. Actor Allu Arjun and businessman Dr. Sanjiv Goenka each obtained interim orders in April 2026 in cases that specifically grappled with AI-generated content, with the Court in Goenka’s matter drawing a notable line between legitimate parody or caricature and unlawful exploitation. A well-known playback singer secured an ex parte ad-interim injunction against AI voice cloning, unauthorised merchandise, false event associations and infringing domain names — with the Court noting the issue “shocked the conscience” given how exposed performers are to commercial exploitation by AI tools. And, as covered in our previous post, cricketers Abhishek Sharma and Yuvraj Singh both secured similar relief in July 2026.
Add to that Hrithik Roshan, Katrina Kaif, Anil Kapoor and, in an earlier and widely cited matter, Amitabh Bachchan — whose case is often described as the starting point for this line of jurisprudence — and the picture is unmistakable: this is not a niche entertainment-law issue. It spans film, sport, business, religion, journalism, politics and music.
The Legal Gap
Here is the uncomfortable truth for anyone advising on this area: India has no dedicated statute governing personality rights or deepfakes. The Copyright Act protects creative works, not a person’s face or voice. The Trade Marks Act protects brand names and logos, not human identity. What exists instead is a multilayered improvisation — constitutional privacy rights, the common-law tort of passing off, scattered IT Rules provisions on intermediary takedown obligations, and criminal law provisions on impersonation — that courts are stitching together, case by case, into something that functions like a personality rights regime without ever being formally legislated as one.
Legal commentators tracking this trend have pointed out that the resulting body of law, while protective, remains scattered: there is no single, consolidated statement of what exactly is protected, for how long, against whom, and with what remedies beyond interim injunctions. That has real consequences. It means the boundaries of “permissible” AI-generated content involving real people — parody, commentary, satire, fan art, brand campaigns “inspired by” a celebrity’s style — are being drawn informally, order by order, rather than through a published statutory test.
Also Read: Personality Rights in the AI Era.
Right to Publicity vs Right to Privacy: What’s the Difference?
The terms get used almost interchangeably in Indian reporting on these cases, but they protect different interests. The right to publicity is the commercial slice — a person’s ability to control and profit from the commercial use of their name, image, voice or likeness; this is what stops unauthorised merchandise, fake endorsements, or an AI chatbot trading on a celebrity’s persona. The right to privacy is the dignitary slice — the right not to have your identity exposed, distorted or exploited against your will, rooted in the Supreme Court’s recognition of privacy as a fundamental right under Article 21. A straightforward unauthorised-merchandise case typically runs on the publicity right alone; a deepfake case, by contrast, usually runs on both at once, because it simultaneously misappropriates commercial value and inflicts a dignitary harm. Understanding which right (or both) is in play matters practically, because it shapes which remedies and which legal doctrines a claim can draw on.
Deepfake Liability for Businesses: What Brands and AI Tool Builders Should Take
This is where the story stops being about celebrities and starts being about anyone running a marketing function, an AI product, or a content operation.
- Marketing and creative teams using AI image or video tools need a clearance process before any output that resembles, references or is trained on a real, identifiable person — even a private individual, and especially anyone with public recognition — goes anywhere near a public campaign. The Sanjiv Goenka order’s careful distinction between parody and unlawful exploitation is instructive: courts are not blanket-banning commentary or satire, but the margin for error is thin and fact-specific.
- E-commerce sellers and merchandise platforms are squarely in the firing line. Nearly every order in this wave — Zinta, Yuvraj Singh, Sharma, the playback singer’s case — included unauthorised merchandise as a distinct head of complaint, separate from the deepfake content itself. Platforms hosting third-party sellers should be auditing listings that use a real person’s name or likeness without a documented licence.
- AI tool builders and deployers, particularly anything offering voice cloning, face-swapping, or chatbot personas, should treat “does this touch a real, identifiable person’s attributes” as a standing compliance checkpoint, not an afterthought. The voice-cloning order’s language — that AI voice cloning of a performer’s attributes violated personality rights — is a direct signal that generative tools are not neutral pipes; deploying them against a real person’s identity carries legal exposure for whoever does the deploying, not just the underlying AI provider.
- Public figures and their representatives — and this now clearly includes athletes, executives, and professionals with any public profile, not only entertainers — should consider proactive registration and documentation strategies: trademarking a signature name or catchphrase where possible, maintaining a documented record of one’s public image and voice for evidentiary purposes, and having counsel ready to seek a John Doe order the moment infringing content surfaces, since courts have repeatedly shown they will move quickly on well-evidenced applications.
The Unresolved Question
What none of these orders yet answer, because none of them are final judgments, is where the outer boundary lies: how long does a personality right last, does it survive death, how does it interact with the fundamental right to free expression, and what happens when a genuinely transformative AI work draws on a public figure’s persona without being deceptive or commercially exploitative? Legal experts following this trend have called for consolidated guidelines precisely because the current approach — protective, fast, but scattered — leaves those questions open on a case-by-case basis.
Until Parliament (or the courts, cumulatively) settle that boundary, the safest operating assumption for any brand, platform or AI deployer in India is this: if your content, product or campaign uses a real, identifiable person’s name, image, voice or likeness without documented consent, you are operating inside a zone the courts have shown a strong and consistent willingness to restrain — regardless of whether a specific statute names your exact conduct.
Building an Internal Clearance Process
For businesses that regularly produce marketing, editorial or AI-generated content, the most useful response to this trend is not a one-off legal opinion but a repeatable internal process. That typically means a short clearance checklist applied before any content referencing a real, identifiable person goes live: has the person (or their authorised representative) provided documented consent for this specific use; if the content is commentary, satire or parody, does it stop short of implying endorsement or generating a commercial product; and if the content was AI-generated or AI-assisted, is there a record of what tool was used and what real-world reference material, if any, informed the output. None of this needs to slow down a marketing team significantly once it becomes routine — the businesses that get caught out tend to be the ones treating each piece of content as a one-off creative decision rather than running it through any consistent screen at all.
Trademark and Personality Rights in India: Why Registration Strategies Still Matter
A recurring question from public figures and their managers on how to protect personality rights in India is whether trademarking a name, signature phrase or nickname adds anything meaningful given that courts are already granting protection through privacy and passing-off doctrines. The honest answer is that it adds a second, independent line of defence: a registered trademark gives its owner a more predictable, faster enforcement route (including against unauthorised merchandise, which has featured in nearly every case discussed above) than relying solely on a fact-specific personality rights claim built afresh each time. For public figures with durable commercial value in a specific name, catchphrase or logo-like signature, pursuing trademark registration alongside — not instead of — a broader personality rights strategy is worth the modest additional cost.
