Bombay High Court orders takedown of Samantha Ruth Prabhu deepfakes

The Bombay High Court on Thursday ordered the takedown of deepfakes and other unauthorised uses of actor Samantha Ruth Prabhu's name, image and voice, and barred further use of those traits without consent. Justice Madhav Jamdar granted the protection as ad-interim relief in a personality-rights suit. The matter is listed again on 10 December.

Prabhu's petition names 18 defendants, among them technology platforms, developers of artificial-intelligence tools, media entities, e-commerce sites and government bodies. Her counsel, Rashmin Khandekar, pointed to morphed photographs, synthetic voice clips, obscene video and interactive chatbots that used her likeness. Jamdar held that the material before the court showed, on a first reading, a breach of her personality and publicity rights through non-consensual deepfakes and morphed images. The order restrains publication of that material and directs its removal.

The suit asks for more than a temporary block. Prabhu wants a permanent injunction, a John Doe order that would cover infringers not yet identified, and damages. Thursday's order does not decide those. It keeps the content down, and the defendants bound, until the December hearing. An ad-interim order of this kind is the court's way of stopping the spread while pleadings are exchanged. It is not a finding on the final facts.

Indian courts have spent the last two years building a line of these cases. Actors have asked Delhi and Bombay benches to treat a face, a voice and a name as rights that survive the clip they were taken from. The legal hook is publicity and personality, tied by the lawyers to privacy and dignity under the Constitution. Jamdar's order sits in that line. What is specific to this file is the mix of targets: not only a website hosting a clip, but tool-makers whose models can generate a new clip, and shops alleged to have used the likeness to sell goods.

A John Doe prayer, if granted in December, would let Prabhu's lawyers serve future hosts without a fresh suit for every URL. That is the remedy plaintiffs in these cases actually use. A damages claim is harder. Platforms will argue they took content down on notice. Tool-makers will argue that a general model is not a use of one actor's face. The court has not reached either argument. It has reached the narrower point that the examples Khandekar placed on record were enough to freeze the identified uses.

The practical effect this week is a set of takedown directions against the named defendants, and a ban on new unauthorised use of the name, image and voice, including AI-made deepfakes. Viewers will still find copies on services outside the order. The suit's answer to that is the John Doe plea, which is not yet granted. Until 10 December the binding text is Jamdar's interim order, the list of 18 defendants, and the categories of material the judge described: morphed images, synthetic voice, obscene video, and chatbots built on her likeness.

Prabhu has not, in the public record of Thursday's hearing, put a number on the clips. The court's finding is qualitative. It saw enough to treat the use as a prima facie wrong. The December date is when the defendants answer, and when the gap between a temporary order and a permanent one gets argued.