Who owns your face? India's legal battle over AI and personality rights

On a recent morning in Pune, 68-year-old Rajesh Sharma settled into his favourite chair after breakfast, opened his phone and began scrolling through social media videos. One clip caught his attention. A well-known Bollywood actor, someone Sharma had admired for years, looked straight into the camera and passionately recommended an investment scheme promising extraordinary returns.

The actor’s expressions were familiar. The voice sounded unmistakably genuine. Even the subtle mannerisms seemed authentic.

Sharma clicked the link. Only later did he discover that the actor had never recorded the video. The endorsement did not exist. It had been manufactured using artificial intelligence.

Sharma did not transfer any money. Many others have not been as fortunate. Across India, people are increasingly being tricked by AI-generated videos, cloned voices and fabricated endorsements that blur the line between reality and fiction. The technology no longer merely edits photographs or alters voices. It can recreate an entire person.

That transformation has triggered a quiet but consequential legal battle. Over the past few years, several celebrities have moved courts seeking protection over something more personal than copyrights or trademarks: their own identities. Amitabh Bachchan, Rajinikanth, Anil Kapoor, Jackie Shroff, Arijit Singh, Sonakshi Sinha, Asha Bhosle and cricketer Gautam Gambhir are among those who have sought judicial intervention against unauthorised use of their names, faces, voices and signature expressions. Their concerns are no longer confined to misleading advertisements or counterfeit merchandise. Increasingly, they concern AI-generated videos, voice clones, manipulated interviews and fabricated endorsements.

At first glance, these cases appear to concern celebrities protecting their commercial value. Faces sell products. Voices build trust. But reducing personality rights litigation to a dispute over endorsements misses the deeper shift unfolding beneath the surface. The courtroom battles are becoming a test case for one of the defining questions of the artificial intelligence era: if technology can create a convincing version of you without your knowledge or permission, who really owns your identity?

That question extends well beyond film stars. A schoolteacher can be made to appear delivering remarks she never uttered. A bank manager’s voice can be cloned to deceive customers. A young woman’s photographs can be transformed into explicit content without her consent. The democratisation of AI has also democratised impersonation. Celebrities have become the first line of defence, not because they are uniquely vulnerable, but because they possess the resources to fight back.

Actor Vivek Oberoi believes the public has misunderstood what these legal battles are really about. For him, personality rights are not merely intellectual property. They represent an individual’s autonomy over how their identity is used. “Even if it is an individual who is not a celebrity, they still have the right to consent,” he says. “I don’t want malicious lies being spread about me, being monetised for personal use, and content that people are creating and leveraging to make money at the cost of the integrity of my brand or my ethics or my core values.”

His argument reflects a profound technological shift. For centuries, identity was inseparable from the individual. A person’s face belonged to them because nobody else could convincingly reproduce it. Artificial intelligence has altered that almost overnight. Freely available software can generate realistic videos from a handful of photographs. Voice cloning tools require only a few seconds of audio. The barriers that once protected identity—expensive studios and teams of visual effects artists—have collapsed, with the work now possible on an ordinary laptop.

The implications stretch beyond entertainment. Financial fraudsters increasingly rely on AI-generated impersonation to deceive investors. Cybercriminals clone voices to bypass verification systems. Political misinformation campaigns deploy manipulated videos. Online extortion rackets exploit deepfakes to manufacture compromising content. The victim need not be famous. In many ways the risk is greater for ordinary citizens, who lack lawyers capable of obtaining emergency injunctions.

Oberoi’s own appreciation of these dangers is shaped partly by personal experience. He recalls a business dispute in which unproven allegations spread rapidly across digital platforms and soon acquired a life of their own. “It impacted me very negatively,” he says. A reputation built over decades can be undermined within hours. Artificial intelligence threatens to accelerate that exponentially, particularly when deployed against vulnerable individuals.

Unlike copyright, patents or trademarks, personality rights have no comprehensive statutory framework in India. Judges have instead assembled protections through constitutional guarantees, intellectual property law and common law principles. The foundation was laid in 1994, when the Supreme Court, in R. Rajagopal v State of Tamil Nadu, recognised a right to privacy and held that aspects of personal life cannot be commercially exploited without consent. High courts subsequently expanded that reasoning to acknowledge that names, likenesses and other distinctive attributes carry commercial value deserving protection.

That evolution has accelerated in the age of artificial intelligence. A defining moment came in 2023, when the Delhi High Court granted actor Anil Kapoor sweeping protection against unauthorised commercial use of his name, image, voice, signature expressions and even his catchphrase. The court acknowledged that personality extends beyond physical appearance, and that gestures, vocal characteristics and mannerisms can all become legally protectable. Subsequent orders involving Bachchan, Shroff, Singh and Sinha reinforced that the law is adapting to digital realities even where legislation has not.

But legal experts warn that courts are essentially building the aircraft while flying it. “There is no single law governing personality rights in India,” says Rodney D. Ryder, founding partner of Scriboard and one of the country’s leading technology lawyers. “What exists today is an amalgamation of constitutional protections, copyright law, trademark law, passing off principles and judicial precedents.” It is a patchwork never designed for an era in which AI can manufacture convincing digital replicas.

Traditionally, copyright protects creative works, trademark safeguards commercial identifiers, privacy protects personal autonomy and passing off prevents misleading consumers. Artificial intelligence slips comfortably through the gaps separating all four. An AI-generated voice clone may not copy an existing recording; it creates an entirely new voice that merely sounds identical. A synthetic image may not reproduce any single photograph but assemble a new face from hundreds of publicly available images. “The challenge is not merely copying content,” Ryder says. “AI can recreate an individual’s identity without directly copying any single copyrighted work.” The irony, he notes, is that the more visible a person becomes online, the easier they become to imitate.

For celebrities, that raises commercial concerns. But Ryder argues the real vulnerability lies elsewhere. Ordinary citizens usually cannot establish the commercial goodwill required for passing off claims, rarely hold registered trademarks over their names, and often lack the resources to litigate. “The law must be applicable to ordinary citizens, not only celebrities,” Ryder says. “Identity protection cannot remain dependent on commercial fame.” A teacher’s voice can be cloned to deceive parents. A doctor’s image can be manipulated into endorsing fake medicines. In many such cases there is no commercial loss to quantify. The injury is reputational and psychological.

The Information Technology Act criminalises identity theft and cheating by personation. The Digital Personal Data Protection Act places consent at the centre of data processing. Copyright and trademark law protect narrower slices of the problem. Collectively, they offer partial remedies. Individually, none was drafted with generative AI in mind. “The current framework is largely inadequate to address AI-generated impersonation comprehensively,” Ryder says. “It remains fragmented and reactive rather than preventive.”

Judges have responded with procedural innovation, chiefly the John Doe injunction, which allows courts to act against unknown defendants before their identities are discovered, and dynamic injunctions, which let plaintiffs seek removal of future copies of infringing material without returning to court each time. Ryder believes the next step is legislative: a dedicated personality rights law defining protected identity, recognising consent as the governing principle, creating fast track takedown mechanisms, introducing statutory damages, and distinguishing malicious impersonation from satire and legitimate artistic expression.

For every actor with an army of lawyers, there are thousands of ordinary Indians with neither the resources nor the legal awareness to fight back. For Vivek Goyal, chief operating officer of digital brand protection firm LdotR, that is precisely why personality rights can no longer be viewed as a celebrity issue. “AI has dramatically lowered the cost and increased the scale of impersonation, making every individual’s face, voice, image and online identity a potential target for misuse,” he says.

Deepfake technology has become faster, cheaper and more accessible, and criminal syndicates have embraced it, combining fabricated videos with phishing campaigns, fake websites and cloned social media accounts. “To an average user, the entire ecosystem may appear authentic,” Goyal says. Even biometric authentication is beginning to face synthetic attacks. Identity itself is becoming hackable.

Governments elsewhere are responding. In the United States, the right of publicity allows individuals to control commercial use of their identities, and Tennessee’s ELVIS Act specifically addresses AI-generated voice cloning. The European Union’s AI Act emphasises transparency and consent, requiring certain AI-generated content to be labelled. China has introduced regulations requiring synthetic media to be identifiable. India is only beginning that conversation, and even strong judicial orders run into a practical obstacle: content removed from one platform reappears elsewhere within minutes, often from servers beyond Indian jurisdiction. “The next phase of personality rights enforcement will not be fought only in courtrooms,” Goyal says. “It will be fought across domains, web sites, social platforms, marketplaces, AI applications and digital channels where impersonation occurs.”

Even Oberoi, among the most vocal advocates of stronger personality rights, accepts that legal protection cannot become an excuse to suppress legitimate criticism. “There is a harmless kind of aspect,” he says. “Somebody making fun, somebody poking satire, jokes, that is okay. That is a hallmark of what a healthy democracy is.” The distinction, he argues, lies in intent: speech that masks “a malicious intent to manipulate the truth, to deceive, to trick, to malign” should not receive the same protection.