Introduction
The proliferation of generative AI technologies capable of producing hyper-realistic synthetic media has created a new category of intellectual property harm that sits at the intersection of trademark law, personality rights, consumer protection, and platform liability. Deepfake advertisements, generated using AI systems trained on images and videos of prominent celebrities, have been deployed in India on an alarming scale to promote fraudulent investment schemes, counterfeit health products, and fake financial services. The individuals whose likenesses have been misappropriated in these campaigns include Sachin Tendulkar, Virat Kohli, Mukesh Ambani, Amitabh Bachchan, and several prominent politicians and media figures.
The legal tools available to combat this phenomenon in India are inadequate in both design and execution. The Trade Marks Act 1999 addresses trademark infringement and dilution of well-known marks, but the connection between a celebrity’s image and a trademark is indirect and requires jurisprudential development. The common law passing off action, extended through decisions such as Titan Industries v. Ramkumar Jewellers (Delhi HC, 2012) to protect personality rights, provides a cause of action but does not create structural deterrence against the scale at which deepfake advertisement campaigns are deployed. The IT Rules 2021 impose takedown obligations on intermediaries but do not mandate proactive detection of deepfake content, and enforcement against foreign platforms hosting deepfake advertisements is practically limited.
This article examines the legal framework available in India for combating AI-generated brand impersonation and deepfake advertisements, analyses the specific doctrinal challenges posed by this phenomenon, and proposes legislative reforms including a standalone Personality Rights Act and platform-specific deepfake liability rules that would more effectively protect individuals and brands against this rapidly growing form of harm.
Legal Framework
Trade Marks Act 1999: Infringement and Dilution
Section 29 of the Trade Marks Act 1999 defines trademark infringement to include use of a registered mark in the course of trade in relation to goods or services identical or similar to those for which the mark is registered, where such use is likely to cause confusion or is likely to be taken as indicating a connection between the goods or services and the trademark owner. The application of Section 29 to deepfake advertisements requires establishing that the celebrity’s name, image, or voice, or some distinctive aspect of their public persona, constitutes or is associated with a registered trademark.
Many prominent Indian celebrities have registered their names and, in some cases, their signatures and distinctive images as trademarks. Amitabh Bachchan, for example, has registered his name, voice characteristics, and distinctive image elements as trademarks in multiple classes. For such registrations, a deepfake advertisement that uses these registered marks in connection with goods or services in the relevant class constitutes infringement under Section 29. The advertisement that uses Bachchan’s AI-generated likeness to promote a fraudulent investment scheme uses his protected mark in connection with financial services without authorisation, satisfying the basic elements of infringement.
Section 11 of the Trade Marks Act protects well-known marks against dilution, which includes both “blurring” (weakening the distinctiveness of the mark through unauthorised use in relation to dissimilar goods or services) and “tarnishment” (harming the reputation of the mark through association with inferior or disreputable products or services). Deepfake advertisements associating a celebrity’s image with fraudulent schemes or counterfeit health products cause precisely this kind of tarnishment, regardless of whether the goods or services promoted are similar to those associated with the celebrity’s registered trademarks.
Personality Rights Under Common Law Passing Off
The most important judicial development in Indian personality rights law for the current purpose is the Delhi High Court’s decision in Titan Industries Ltd. v. Ramkumar Jewellers (2012). In that case, the Court held that a jewellery company that used images of Amitabh Bachchan and Jaya Bachchan, brand ambassadors for a competing jewellery brand, in its own promotional materials without authorisation committed passing off. The Court held that the commercial value inherent in a celebrity’s image, including the endorsement goodwill built through the celebrity’s association with specific brands, is a property right that the law of passing off protects.
The Titan decision established three elements for personality rights passing off: the celebrity must have a recognisable “image” or persona with commercial value; the defendant must have used that image in a way that represents, falsely, that the celebrity endorses or approves of the defendant’s goods or services; and that misrepresentation must cause or be likely to cause damage to the celebrity. Deepfake advertisements satisfying all three elements: they appropriate the celebrity’s image with commercial value, they falsely represent the celebrity’s endorsement of the advertised product, and they cause reputational damage and loss of endorsement value.
The limitation of passing off as a remedy for deepfake advertisements is practical rather than doctrinal: obtaining an injunction against a specific deepfake advertisement requires identifying the infringer and commencing proceedings in time to prevent the advertisement from causing harm, a process that is typically too slow relative to the viral spread of deepfake content on social media platforms. By the time an injunction is obtained, the advertisement may have been viewed millions of times and the reputational damage may be irreversible.
IT Rules 2021 and Intermediary Obligations
The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules 2021, issued under the IT Act 2000, impose obligations on social media intermediaries to take down content violating specified categories, including content that impersonates another person or is “patently false and untrue or misleading in nature.” Rule 3 requires intermediaries to “publish rules and regulations, privacy policy and user agreement” informing users about prohibited content, and to establish a grievance redressal mechanism. Significant social media intermediaries (with more than five million registered users in India) must additionally appoint a grievance officer, a nodal contact person, and a Chief Compliance Officer.
The Rules do not impose any obligation on intermediaries to proactively identify deepfake content or to use AI detection tools to prevent deepfake advertisements from being published. The obligation is reactive: platforms must take down content once notified. This reactive model is structurally inadequate for deepfake advertisements, which can reach millions of viewers within hours of publication, causing harm that cannot be remedied by subsequent takedown.
Judicial Developments
Celebrity Deepfake Litigation in India: 2023-2025
The volume of deepfake advertisements exploiting Indian celebrities’ likenesses grew substantially in 2023 and 2024. Several celebrities, including Sachin Tendulkar and Aamir Khan, publicly complained about deepfake videos promoting online gambling and investment fraud schemes, triggering criminal complaints to the Cyber Crime Cells of state police forces and requests to platforms to take down the content. These complaints did not, for the most part, generate civil judgments because they were addressed through platform takedown mechanisms rather than litigation.
The Delhi High Court’s 2024 order in Anil Kapoor v. Sukesh Chandrashekhar and Others addressed personality rights in a context involving the unauthorised commercial use of Kapoor’s name and image in merchandise and promotional material. While the case did not specifically concern AI deepfakes, the Court’s broad articulation of personality rights protection, including the right to control commercial exploitation of one’s name, image, voice, and distinctive persona, provides a legal foundation that courts can apply to deepfake cases.
The Madras High Court has addressed related questions in cases concerning morphed images and digitally manipulated photographs published on social media. These decisions, while not specifically addressing generative AI deepfakes, establish that digital manipulation of a person’s image for commercial or defamatory purposes is actionable under passing off and tort principles.
Contemporary Issues and Analysis
The Scale Problem: Generative AI and Viral Deepfakes
Traditional personality rights and trademark dilution litigation is designed for adversarial disputes between identifiable parties over specific uses. It is not designed for an environment in which thousands of individual actors, operating pseudonymously through foreign platforms, can generate and distribute deepfake content at scale using commercially available AI tools. The Indian government’s Cyber Crime reporting portal (cybercrime.gov.in) received tens of thousands of complaints related to deepfake and morphed image content in 2023 and 2024, but the conversion rate from complaint to prosecution was negligible.
The structural problem is that existing legal remedies are calibrated for the wrong scale. Individual injunctions against individual infringers are obtainable but inconsequential when any individual who acts on an injunction can be replaced by thousands of others using the same or similar AI tools. The appropriate response requires a combination of platform-level obligations, technical standards for deepfake labelling and detection, and collective enforcement mechanisms that can operate at the scale of the problem.
The Endorsement Economy and Measurement of Harm
Indian celebrities earn substantial income from brand endorsement contracts, which typically include exclusivity provisions limiting the celebrity’s ability to endorse competing products. Deepfake advertisements that falsely portray celebrities as endorsing products create multiple categories of harm: they damage the celebrity’s reputation by associating them with fraudulent or inferior products; they dilute the value of the celebrity’s legitimate endorsements by creating consumer confusion about which endorsements are authentic; and they may trigger breach of exclusivity provisions in legitimate endorsement contracts if consumers mistakenly believe the deepfake endorsement is genuine.
Quantifying these harms for the purpose of damages assessment is methodologically challenging. Survey-based evidence of consumer confusion, valuation evidence of endorsement market rate reduction, and evidence of actual damage to existing endorsement contracts are the most probative categories, but assembling this evidence for litigation is expensive and time-consuming.
Comparative and International Perspective
California’s AB 602 and AB 2602 (2023) represent the most direct legislative response to AI deepfakes in the entertainment context. AB 2602 creates a right of digital replica consent: before creating a digital replica of a performer’s voice or likeness for use in audiovisual work, the contracting party must obtain explicit written consent. The legislation applies to new contracts but also invalidates provisions in existing contracts that would require digital replica consent by implication rather than explicit agreement.
The EU’s AI Act (Regulation (EU) 2024/1689), which entered into force in 2024, requires that synthetic media created by AI systems, including deepfake videos, be transparently labelled as AI-generated. This transparency requirement does not prevent deepfake creation but creates a legal basis for enforcement against unlabelled deepfakes and supports consumer recognition of synthetic media. The EU’s approach of transparency regulation, rather than prohibition, is flexible and technologically neutral.
WIPO’s Intergovernmental Committee on Intellectual Property and Genetic Resources, Traditional Knowledge and Folklore has not specifically addressed AI-generated deepfakes, but WIPO’s 2024 “Issues Paper on Artificial Intelligence and IP Policy” identifies personality rights as an area requiring international coordination as AI technologies advance.
Practical and Policy Implications
For celebrities and brands, the immediate practical priority is comprehensive trademark registration of name, image, and distinctive persona elements across all relevant classes, combined with active monitoring of digital platforms for infringing content using commercial brand monitoring services. Where deepfake advertisements are detected, the priority should be submitting takedown requests to platforms under their community standards and reporting processes, supplemented by statutory notices under IT Act Section 79, before commencing litigation.
For brand advertisers and marketing agencies, the legal risk of deepfake technology in the advertising context is not only the risk of their own brand being impersonated but also the risk that AI-generated advertising content they produce may inadvertently incorporate reproductions of protected likenesses or trademarks. Robust AI content governance policies that include provenance checking for AI-generated images are now a necessary element of commercial risk management.
Suggestions and Reforms
India requires a standalone Personality Rights Act that codifies the right of individuals, particularly those with commercially significant public personas, to control commercial exploitation of their name, image, voice, likeness, and distinctive persona elements. The Act should provide for both civil remedies (injunctions, damages including statutory damages, accounts of profits) and criminal penalties for commercial exploitation without consent. It should create a presumption of consumer confusion and reputational harm in cases of deepfake advertisements, relieving claimants of the burden of proving these elements individually.
The IT Rules 2021 should be amended to impose proactive deepfake detection obligations on significant social media intermediaries. Platforms with over five million Indian users should be required to implement AI-based detection systems for deepfake content, to label AI-generated video content transparently, and to prevent the publication of deepfake advertisements without verification of the consent of persons depicted. The implementation timeline should be phased to allow platforms adequate time to develop technical compliance mechanisms.
The Consumer Protection Act 2019 should be amended to specifically address deepfake advertisements as a category of misleading advertisement prohibited under Section 2(28). The Central Consumer Protection Authority should be empowered to impose substantial fines on advertisers who use deepfake technology to create false endorsements, with liability extending to advertising agencies that knowingly facilitate such campaigns.
Conclusion
AI-generated deepfake advertisements represent a qualitatively new form of brand impersonation and personality rights violation that India’s existing legal framework, designed for a world of physical and traditionally digital misappropriation, is inadequately equipped to address. The combination of trademark law, passing off doctrine, and IT Rules takedown obligations provides some legal tools, but their effectiveness is limited by speed, scale, and enforcement challenges that require systematic legislative response.
The reforms proposed in this article, a Personality Rights Act, proactive platform obligations, and consumer protection enforcement, represent a coherent legislative strategy for addressing deepfake brand impersonation at the scale and speed the problem demands. India has a compelling economic and cultural interest in developing this framework: its entertainment industry, its celebrity endorsement economy, and its consumer markets are all materially harmed by the unchecked proliferation of AI-generated brand impersonation, and legislative action is both justified and overdue.
