NEW DELHI, INDIA — A Delhi High Court order gives veteran television journalist Rajat Sharma and India TV a continuing, fast-track process for seeking removal of new AI-generated deepfakes and other material that misuses Sharma’s identity or the broadcaster’s protected content.
The Aug. 24 order permanently restrains specified defendants from misusing Sharma’s name, face, voice and other personality attributes, including through artificial intelligence-generated deepfakes. It also bars those defendants from infringing India TV’s trademarks and copyrighted programming.
What the court ordered
The ruling creates a continuing response process for qualifying new material. Sharma and India TV may send reported links to Meta or Google. The platforms must acknowledge the request within 24 hours and, after independently reviewing the material under their own policies, act within another 36 hours. If the content remains online, the plaintiffs may return to the court.
The permanent injunction applies to defendants who were served but did not file defenses. The decision therefore was not the product of a fully contested trial. Meta and Google were not found liable, and the injunction itself does not run against the platforms. The court also declined to declare the India TV marks legally “well-known.”
Why radio should pay attention
The case shows how a media company can combine personality rights, trademark protection and program copyright to challenge fake endorsements and fabricated programming. For radio, the closest parallel is an unauthorized cloned voice used to impersonate a host, sell a product or create a fake station segment.
Radio groups may want to review whether talent agreements clearly address voice and likeness rights, whether station and program names are properly protected, and whether original shows and airchecks are preserved well enough to document ownership. A ready-to-use platform reporting process could also shorten the response time when a fraudulent clone appears.
Important limits
This is an Indian court decision and is not binding in the United States. It concerns unauthorized outputs and impersonation, not whether copyrighted material may lawfully be copied to train an AI model. It also creates no new U.S. obligation or liability for a station merely testing an AI tool.
The strongest takeaway for broadcasters is practical: recognizable talent and trusted station brands can be used to make synthetic content appear authentic, so contracts, registrations, archives and platform-response procedures should be in place before a deepfake circulates.
Primary document: Delhi High Court order in Rajat Sharma & Anr. v. Tamara Doc & Ors.
This article provides general information and is not legal advice.