India has no law that says, in so many words, "you cannot clone a person's voice or face without consent." And yet, through 2026, its courts have been acting as if one exists — building a real body of personality-rights protection one injunction at a time.
The cases stacking up
The pattern is consistent and increasingly frequent:
- The Bombay High Court ordered the takedown of AI deepfakes and sexually explicit synthetic content involving actress Shruti Haasan.
- The Delhi High Court restrained similarly explicit synthetic content involving Meenakshi Chaudhary.
- Actor Manoj Bajpayee moved the Delhi High Court against AI voice generators, deepfakes, and merchandise trading on his name and likeness — seeking ₹2 crore in damages.
- The Delhi High Court also ordered the takedown of AI-generated content infringing the personality rights of Drishti IAS's founder, with the bench reportedly describing certain AI tools as "worse than a gun" in their capacity for reputational harm.
The legal tools courts are actually using
Without a dedicated personality-rights statute, judges are stitching protection together from existing law: the Copyright Act, the Trade Marks Act, the IT Act, and constitutional rights to dignity and privacy. Courts are granting injunctions against voice cloning, deepfakes, and synthetic endorsements case by case — effectively legislating from the bench while Parliament stays silent on a standalone framework.
The regulatory piece: IT Rules amendment
On February 10, 2026, the government amended the Information Technology Rules, 2021 specifically to strengthen oversight of synthetically generated content. The amendment introduces requirements for identifying and labelling synthetic audio-visual content, traceable metadata, and faster takedown processes — a direct regulatory response to the same deepfake problem the courts have been handling through injunctions.
Why this is now a business risk, not just a celebrity problem
Any business using AI-generated voice, likeness, or synthetic media — in advertising, marketing, virtual assistants, or content production — now operates against a backdrop of real, court-tested liability, even though no single statute defines the boundaries. The combination of an expanding body of case law and a labelling-and-metadata regulatory regime means "we didn't think it counted as a violation" is a weaker defence with every new ruling.
The practical takeaway: get explicit consent before using anyone's voice or likeness in synthetic content, label AI-generated media clearly, and watch this space — a standalone personality-rights or deepfake statute is a plausible next step once the case law matures further.