36 Hours to Take It Down: What the Delhi High Court's Rajat Sharma Deepfake Ruling Means for Personality Rights in the AI Era

In Rajat Sharma & Anr. v. Tamara Doc & Ors. (2026 LLBiz HC(DEL) 954, CS(COMM) 1147/2024), an order dated 24 August 2026, Justice Jyoti Singh of the Delhi High Court granted a permanent injunction protecting journalist Rajat Sharma's personality and publicity rights against AI-generated deepfake videos that used his image, voice, and mannerisms without authorization. The Court also directed Meta and Google to take down infringing content within a fixed 36-hour window once notified. Here is the case, what personality rights actually protect, and what the ruling means for anyone whose face or voice could end up in an AI-generated video without their consent.

Nearly Three Decades of a Recognizable Voice, Then a Fake Version of It

Rajat Sharma has hosted Aap Ki Adalat for 29 years, making his face and voice among the most recognizable in Indian television journalism. That recognizability became the problem. Multiple YouTube channels were found circulating videos built using artificial intelligence to create distorted, modified, and modulated versions of Sharma's image, voice, and personality traits, without his authorization, effectively putting words in his mouth that he never said. Sharma, along with INDIA TV, filed suit covering not just this misuse of his persona but also unauthorized use of INDIA TV's registered trademarks and related copyright violations.

What "Personality Rights" Actually Protect

Personality rights give an individual control over the commercial use of their name, likeness, image, voice, signature, and other attributes that make up their public persona. They are distinct from copyright, which protects original creative works, and from trademark, which protects brand identifiers like logos and names. In India, personality rights have developed through a combination of the right to privacy recognized under Article 21 of the Constitution and common law passing-off principles, and Indian courts have granted this protection with increasing frequency in recent years, largely in cases involving film and television celebrities. This ruling confirms the protection extends just as firmly to a journalist whose persona has commercial and reputational value built over decades of public-facing work, not only to film stars.

Why AI Deepfakes Raise the Stakes

Traditional impersonation or crude photo editing is one thing. AI-generated deepfakes are another, because they can produce highly convincing fake video and audio that puts synthetic words directly into a real person's mouth, in their own voice and mannerisms, at a scale and realism that older forms of misuse could never achieve. The Court's language describing "distorted, modified, and modulated versions" of Sharma's image and voice reflects a judiciary treating AI manipulation of a real person's persona as squarely within the scope of personality rights protection, rather than as some novel category requiring entirely new legal reasoning.

What the Court Ordered

The Court decreed a permanent injunction restraining the defendants from further infringing Sharma's personality rights and INDIA TV's trademark registrations. It went further by building an actual enforcement mechanism into the order. Meta and Google were directed to acknowledge takedown requests within 24 hours and execute removal within 36 hours following their own independent policy review. The order also set up a dynamic protocol allowing Sharma and INDIA TV to submit newly discovered infringing URLs, supported by an affidavit, for expedited action without needing to return to court for every new violation. If the platforms fail to act within these timeframes, the plaintiffs retain the ability to seek further directions from the Court. Separately, Google had already been ordered to disable the identified channels and disclose subscriber and IP information, and two accounts were terminated with two more still under review at the time of the final order.

A Named-Defendant Order With a Twist

Unlike the broad John Doe or Ashok Kumar orders Indian courts sometimes issue to cover unknown future infringers in advance, this decree targeted specifically identified defendants and channel operators who were served with the suit but did not respond, leading to their default. What makes the order function more broadly in practice is the dynamic takedown protocol, which requires ongoing platform cooperation for newly discovered infringing content without forcing Sharma to file fresh litigation every time a new deepfake surfaces.

What This Means for Public Figures and Businesses

Public figures whose recognizable persona could be exploited through AI-generated content now have a concrete judicial template to work from, one that combines personality rights, trademark protection, and platform-takedown mechanisms with actual enforceable timelines rather than vague good-faith commitments. Businesses built around a recognizable spokesperson, founder, or brand ambassador should treat ongoing monitoring for this kind of misuse as a real operational need, since a structured and reasonably fast legal response is now demonstrably available. Platforms, for their part, are on notice that courts are willing to impose fixed takedown windows rather than leaving removal timing entirely to internal community-standards discretion.

This blog is for general informational purposes and does not constitute legal advice. For guidance on personality rights, trademark protection, or responding to AI-generated content misuse, please contact our team.