The Delhi High Court directed X to take down an AI-generated deepfake video purportedly showing Congress MP Shashi Tharoor praising Pakistan's diplomacy. The court also asked Meta to ensure that certain offending reels on Instagram, which it made inaccessible, continue to remain so.
Shashi Tharoor had reportedly moved the court against the repeated publication of deepfake videos purportedly showing him making "politically sensitive" statements. His senior counsel asserted that such content not only tarnished his reputation but also affected India's international standing.
Justice Mini Pushkarna, in an interim order passed on Tharoor's lawsuit, also restrained the misappropriation of the Thiruvananthapuram MP's name, image, distinct voice, "signature oratorical cadence and manner of speaking", "highly refined vocabulary", and other facets of his persona to create and publish any deepfakes, voice-cloned audio, morphed videos, etc., for any commercial, political, or malicious purpose across any physical or virtual medium.
According to news agency PTI, the interim order said Tharoor is a "respected and recognised public figure" who has "exclusive control" over the utilisation of his personality, and misappropriation of any attribute of his personality without his express permission and any consequent harm to his reputation is liable to be restrained.
"It is no more res integra that personality rights/publicity rights are protectable under Articles 19 and 21 of the Constitution of India, 1950. The plaintiff's reputation, goodwill, name, physical appearance/image/likeness, voice, mannerisms, styles, signature oratorical style, and other attributes are uniquely identifiable and associated with the plaintiff. The same form the plaintiff's 'personality', over which the plaintiff enjoys sole and exclusive control," the court stated.
"Defendant no. 1 (Ashok Kumar / John Doe) are restrained from reproducing, misappropriating, or imitating any facet of the plaintiff's persona, including, but not limited to his (i) name, (ii) visual likeness and image, (iii) distinct voice, (iv) signature oratorical cadence and manner of speaking, and (v) highly refined vocabulary, to create, publish, or disseminate any synthetic media, deepfakes, voice-cloned audio, or morphed videos, through the use of AI, Generative AI, Machine Learning, or any other technology, for any commercial, political, or malicious purpose, across any physical or virtual medium. Defendant no. 2 is directed to forthwith take down and block access to the following link as available on its platform 'X'," it ordered.
The court clarified that in the case of further "false, fake and infringing videos", Tharoor shall be at liberty to approach the social media platforms for taking them down.
In the lawsuit, Tharoor said that in or around March 2026, he discovered a "sophisticated, malicious campaign orchestrated by unknown infringers" across social media platforms, maliciously depicting him making politically sensitive statements praising Pakistan.
Tharoor was represented by senior advocate Amit Sibal and the law firm Trilegal.
The lawsuit contended that the unauthorised cloning and exploitation of Tharoor's likeness, voice, and mannerisms infringed his personality and publicity rights and also constituted a serious violation of his right to privacy.
"These infringers have weaponised artificial intelligence (AI) and machine learning to generate hyper-realistic audio-visual deepfakes by cloning the plaintiff's face, voice, vocabulary, and mannerisms. These fabricated videos maliciously depict the plaintiff making politically sensitive statements that he never made," the plea said.
"Crucially, as the plaintiff was actively campaigning for the Kerala Legislative Assembly elections 2026 in March and early April, the disinformation campaign was especially damaging. It was a deliberate attempt to tarnish his patriotic credentials, manipulate public perception, and unlawfully interfere with the democratic electoral process," it added.