How to Protect Your Name and Image From Deepfakes: Personality Rights

How personality rights protect you against a deepfake

A deepfake that uses your face or voice is, in legal terms, an unauthorised use of your identity. Indian courts treat a person's name, image, likeness, voice, signature and distinctive mannerisms as part of their personality, and they protect it from two kinds of misuse: commercial exploitation without consent, and uses that degrade, falsify or misrepresent the person. No single statute says this. The Delhi High Court has built the right from Article 21 of the Constitution (privacy and dignity, recognised in Justice K.S. Puttaswamy v. Union of India, 2017), from the law of passing off, and from the Trade Marks Act 1999 and the Copyright Act 1957.

The turning point for personality rights deepfake cases was Anil Kapoor v. Simply Life India (2023), where the court expressly restrained use of the actor's likeness through artificial intelligence, face morphing and deepfakes, along with his catchphrase. Earlier orders, such as Amitabh Bachchan v. Rajat Nagi (2022), had protected name, voice and image across merchandise, apps and websites. Through 2025 and 2026 a long line of actors, singers, sportspersons, spiritual leaders and broadcasters obtained similar orders against AI voice clones, chatbots and fake video ads.

What this gives you in practice is a set of court orders that move fast. The court can restrain named defendants and unknown persons (often described as John Doe or Ashok Kumar defendants), direct platforms to remove listed links, require them to act on further links you notify, and order disclosure of who runs the offending accounts. The legal basis and the way these orders are framed are explained on our page on personality rights and deepfake cases in the Delhi High Court. This guide is about what you should do, and in what order.

There is a criminal side too. Identity theft and cheating by personation using a computer are offences under Sections 66C and 66D of the Information Technology Act 2000, violation of privacy under Section 66E, and publishing obscene or sexually explicit material under Sections 67 and 67A. Defamation falls under Section 356 of the Bharatiya Nyaya Sanhita (old Sections 499 and 500 IPC). A police complaint and a civil suit can run side by side, and often should.

Do you have to be a celebrity to use these rights?

No, though the strength and legal basis of the claim change. Celebrity personality rights rest mainly on publicity value: a famous face sells products, so using it without a licence takes money that belongs to the person. That is why actors and cricketers get wide injunctions covering merchandise, ringtones, GIFs, stickers and AI voice models. A private person has no commercial goodwill of that kind, so the claim rests on privacy, dignity and, where false statements are made, defamation.

This matters more than people think. Most victims of morphed images and deepfakes are not public figures. They are women targeted by former partners or extortion gangs, students whose photographs are lifted from social media, and professionals whose faces are pasted into fraudulent ads. The misuse of image in these cases is often worse than anything a film star faces, because the content is sexual or the purpose is to cheat people who trust the face they see. Courts have acted for such persons and can mask their names in the case title.

In between sit people with a modest public profile: doctors, chartered accountants, coaches, teachers, YouTubers with a few lakh subscribers, and business owners who appear in their own ads. They have some commercial reputation and a strong dignity claim. For them, evidence of confusion, such as messages from people who believed the fake ad, is often what persuades the judge.

The court also asks if people can actually identify you in the misuse. A blurred likeness or a common name used in an unrelated context may not support an injunction; a clear face, a recognisable voice, or your name paired with your profession usually will.

What you can do before anything goes wrong

Prevention will not stop a determined fraudster, but it changes how quickly you get an order and how readily a judge believes you. Start by making your genuine presence easy to verify. Keep one public page, on your website or a verified profile, listing your official accounts, your agency or manager, and a line saying you do not endorse investment schemes, medicines or apps unless announced there. When a fake appears, that page is the first document in your evidence.

If your name is used in trade (a clinic, a coaching brand, a stage name, a range of products), register it as a trademark in the relevant classes. Registration gives you a statutory infringement claim and lets you sue where you reside or carry on business under Section 134 of the Trade Marks Act. Our guide on protecting a trademark before going to court explains the filing choices. Book the matching domain names as well, because look-alike domains are a common tool for fake endorsements.

Keep the originals of your photoshoots, videos and recordings, with dates and the names of the photographers or studios. Copyright in a photograph or video usually belongs to the person who made it or the person who commissioned it, depending on the contract. If you own the copyright in the source footage used for a deepfake, you have an extra and often simpler ground for takedown. Performers also have rights in their performances under Sections 38 and 38A of the Copyright Act.

Set up some monitoring: search alerts on your name, a reverse image search every few weeks, and a trusted person who tells you when something looks wrong.

Clauses to add to endorsement and content contracts

• Limit the licence. State exactly which images, videos and recordings the brand may use, for which products, in which media and for how long.
• Ban synthetic use. Prohibit the brand and its agencies from creating AI-generated, altered or voice-cloned content of you without your written approval of each item.
• Require removal on expiry. When the contract ends, the brand must take your material off its channels and ask its distributors and affiliates to do the same within a fixed number of days.
• Control training data. Prohibit use of your recordings or images to train or fine-tune any AI model.
• Fix responsibility for leaks. If your raw footage leaks from the brand's side and ends up in a deepfake, the brand must cooperate with your complaints and court case and bear agreed costs.
• Choose the forum. An exclusive jurisdiction clause for Delhi courts, or an arbitration clause seated in Delhi, avoids an argument later about where to sue.

The first 48 hours after you find a deepfake

Do not start by commenting on the video or messaging the uploader. That alerts them, and many will delete the content and the account, taking your best evidence with them, then upload again under a new name. Save first. Record the screen while the video plays, showing the URL, the account name, the view count and the date. Take screenshots of the profile, the comments, and any payment details, phone numbers or links given in the post.

Note where you found it and who told you. If a patient, client or relative was approached by the fake, ask them for screenshots of what they received and a short written account in their own words. Those messages prove confusion and harm, which a judge looks for before granting a wide order. Keep the originals on your phone and do not crop or edit them.

Then report. Use the platform's own form and pick the narrow category that fits (impersonation, non-consensual intimate imagery, synthetic or manipulated media) rather than a general complaint, and note the complaint number. File a complaint on the national cybercrime portal or call 1930, especially if money is being collected in your name or the content is sexual. If someone threatens to publish images unless you pay, do not pay. Payment rarely ends the demands.

Within the same two days, decide whether this is a one-off item or a pattern. One fake profile that the platform removes may need nothing more. Twenty ads across three platforms, a voice clone on an app, or content that returns after every removal points to a civil suit, and the sooner it is filed the less the content spreads.

How to respond to a deepfake, step by step

This is the order usually followed when someone brings a deepfake or a fake endorsement to a lawyer. Several steps run at the same time.

1. Preserve the evidence

   Make screen recordings and screenshots showing the URL, account handle, date and any payment or contact details in the content. Keep a simple table of each item: where it appeared, when you found it and what you did about it.

2. Report to the platform

   Use the specific reporting category and keep the reference number. Under the IT Rules 2021, every large platform must have a Grievance Officer in India, and content impersonating a person or showing intimate images must be removed within a short fixed time of the person's complaint. Amendments notified in February 2026 shortened these windows further and added labelling duties for AI content, so check the current limit.

3. Complain to the cybercrime police

   Report on the national cybercrime portal or helpline 1930 and to the cyber police station of your district, and keep the acknowledgement. If no FIR follows where the content is clearly criminal, the remedies on our page on Delhi Police not acting on a complaint apply.

4. Issue a short public clarification

   Post on your verified accounts that the video or ad is fake and that you endorse no such product. Keep it factual and do not name individuals you cannot identify with certainty, or you may face a complaint yourself.

5. Send a legal notice where a known party is involved

   If a company, agency or app behind the content can be identified, a notice demanding removal and a written undertaking can end the matter. Where the uploaders are anonymous, this step is usually skipped.

6. File the suit with an interim injunction application

   The suit names known defendants, unknown persons, the platforms and, where needed, the Department of Telecommunications and MeitY for blocking. Because urgent interim relief is sought, pre-institution mediation under Section 12A of the Commercial Courts Act is not required.

7. Enforce the order

   Serve the order on the platforms, check compliance, notify new links as they appear, and use the disclosed account details to add the real people behind the content as defendants.

Platforms and police: what usually happens

Platform complaints work well for single items, especially intimate images and obvious impersonation. They work badly against fraud networks running dozens of ads from fresh accounts, against apps hosted outside India that ignore notices, and against content the platform's reviewers treat as borderline, such as a realistic deepfake labelled as parody. If the Grievance Officer refuses or does not reply, the Grievance Appellate Committee under the IT Rules can be approached, but that takes time you may not have.

Police complaints are needed when money is being taken, when the content is sexual, or when you are being threatened. In Delhi, each police district has a cyber police station, and the Delhi Police also has a specialised cybercrime unit (IFSO). A complaint creates a record, and an FIR allows the police to seek account data from platforms. In practice, action against anonymous uploaders based abroad is slow. Do not wait for the investigation to end before protecting yourself in civil court.

A frequent complication is the counter-complaint. Someone you named in a public post files a defamation complaint, or a person who lost money to the fake ad complains against you, believing you were part of the scheme. Keep public statements factual and keep every message from people who were cheated. Our guide on criminal defamation cases over social media posts explains how such complaints are handled, and our page on cyber crime matters before the Delhi High Court covers FIRs, bail and quashing at the High Court stage.

When a court case becomes necessary

Go to court when the misuse is commercial, repeated, spread across platforms, or coming back after removal. Go also when the content is so damaging (a sexual deepfake, a fake confession, a video putting false political or communal words in your mouth) that you cannot wait in a platform's queue. A civil suit gives you what complaints cannot: one order binding all the platforms named, operating against unknown persons, with a way to add new links without filing a fresh case.

The usual reliefs are a permanent injunction against use of your name, image, voice and likeness, including through AI, deepfakes and voice cloning; takedown of listed URLs; suspension of domains; blocking directions for websites; disclosure of subscriber information; and damages or an account of profits. At the first hearing the court considers the interim injunction. Where the evidence is clear and the harm is continuing, an order without waiting for the other side is common, as explained in our guide on ex parte injunctions in IP cases.

When Advocate B.K. Singh prepares a personality rights deepfake suit, he makes sure every prayer matches a category of misuse that is actually in the evidence: if there are fake ads and a voice clone, the prayers say so, with examples annexed for each. Prayers asking the court to stop anything anyone might ever say about the client get narrowed or delayed. Judges grant wide relief readily when the plaint shows exactly what is being stopped and why.

A writ petition is rarely the right tool. Fundamental rights bind the State, and against private uploaders the remedy is a suit. A writ can help where a government authority refuses to act on a blocking request, or where systemic directions to search engines and police are needed, as the Delhi High Court gave in a 2023 case about intimate images that kept resurfacing online.

Which court hears it: Delhi High Court or the district court

Suits by public figures are usually filed on the original side of the Delhi High Court, often as commercial suits because they include passing off and trademark claims, and they are listed before the Intellectual Property Division. That route needs a valuation above the High Court's pecuniary limit, which is Rs 2 crore at present; a proposal to raise it is pending, so check the current limit. Our guide on the Intellectual Property Division of the Delhi High Court explains how that bench works.

A private individual's privacy claim, valued modestly, normally goes to the district court with territorial jurisdiction, such as Saket for South Delhi or Tis Hazari for central and north Delhi. That court can also grant injunctions and takedown directions. Appeals against its orders on interim injunctions come to the High Court.

Delhi's courts have jurisdiction where the plaintiff lives or works here, where the content is accessed and causes harm in Delhi, or, for registered trademarks and copyright, where the plaintiff resides or carries on business (Section 134 of the Trade Marks Act and Section 62 of the Copyright Act). A person in Noida or Gurugram whose reputation is harmed in Delhi can usually sue here. Someone with no real connection to Delhi should not file here just because the court has settled formats for these suits; the defendants will object, and the objection costs months.

Time and cost at each stage (typical ranges, not promises)

Documents and evidence to collect

What a Delhi High Court order usually covers, and how to use it

A typical interim order in these cases restrains the defendants, including unknown persons, from using the plaintiff's name, image, voice, likeness and other attributes through any technology, including AI, machine learning, deepfakes, face morphing and voice cloning, for commercial gain or in a way that harms the plaintiff. It directs named platforms to remove listed URLs within a fixed time, and allows the plaintiff to send further URLs directly to the platform, which must act on them. If a platform disputes that a new link is covered, the plaintiff can return to the court.

Domain registrars are told to lock and suspend domains using the plaintiff's name, and the Department of Telecommunications and MeitY are asked to issue blocking orders for websites that cannot be reached any other way. These dynamic mechanisms came from piracy cases, which our guide on dynamic injunctions against piracy websites describes.

The order is only as good as the follow-up. Someone must watch for new uploads, send links in the format the order requires, record when each platform complied, and report defaults to the court. After disclosure, the persons behind the accounts can be added as defendants, and a damages claim against an identified seller of fake products becomes realistic. Breach of the order is contempt under the Contempt of Courts Act 1971.

Advocate B.K. Singh's practical advice to clients with a live order is to keep one spreadsheet of every URL sent, the date and time, the platform's reply and how long it took to act. When a platform drags its feet, that log, filed with a short application, is what persuades the court to direct compliance within hours rather than days.

What personality rights will not stop

Personality rights do not give anyone a veto over what others say about them. News reports, criticism, reviews, satire, parody, caricature, biographies and art are protected expression under Article 19(1)(a). A stand-up comedian imitating an actor's voice, a news channel using a cricketer's photograph to report a match, or a meme mocking a politician's speech will generally not be stopped. In Digital Collectibles v. Galactus Funware (2023), the Delhi High Court declined to treat the use of cricketers' names and public statistics in an online game as a violation.

Deepfakes test this line. A clearly labelled parody may be protected, while a realistic video that viewers will take as genuine, or that sexualises or defames the person, is not. The court looks at whether the use exploits the person commercially, creates a false impression of endorsement, or degrades them. Labelling something as AI-generated does not by itself make it lawful.

Accurate but unflattering content is a matter for defamation law, if anything, with its defences of truth and fair comment. Our guide on filing a defamation suit in the Delhi High Court covers that route. Trying to use personality rights against fair criticism usually fails, and it can make the story bigger than it was.

If you are a private person targeted with morphed or intimate images

Start with the platform's intimate imagery reporting route and the cybercrime portal, because the rules require faster removal for this category than for almost any other. Ask a trusted friend or family member to help with the screenshots if looking at the content is too painful; you need a record, but you do not have to collect it alone. Keep safe the phone or device the images may have come from, as the police may need it.

If you know who made or shared the content, such as a former partner, the offences of voyeurism and stalking under Sections 77 and 78 of the BNS (old Sections 354C and 354D IPC) and Sections 66E, 67 and 67A of the IT Act may apply, and an FIR is the right first step. If the content keeps resurfacing on new sites, a civil suit for privacy with directions to search engines to de-index the material can help. The court can allow your name to be masked in the case title and in its orders.

Students and young people should tell a parent or teacher early. Extortion gangs rely on shame and delay. A quick complaint, a refusal to pay and a blocked contact end most of these attempts, though some content may still need a court order to stay down.

Mistakes that weaken a deepfake case

• Confronting the uploader first. The account disappears with the evidence, and the content returns under a new name.
• Cropped or edited screenshots. They look tampered with. Keep full-screen captures that show the URL and date.
• Waiting for the police to finish. An investigation and a civil suit can run together; delay lets the content spread and weakens the urgency argument for an interim order.
• Over-broad prayers. Asking the court to stop all mention of you invites free speech objections and delays relief.
• Accusing the wrong person in public. Naming a suspect without proof can lead to a defamation complaint against you.
• Ignoring old endorsement contracts. A brand that still runs your photographs after its licence expired is a contract and passing off case; read the agreement before treating it as a deepfake.
• Filing in a court with no connection to the facts. Jurisdiction objections cost months, so choose the forum on the facts.

Advice on protecting your identity

Most identity misuse can be contained if you save the evidence, report it on the right channels in the first two days, and move to court when it is repeated or commercial. Preparation (a public list of official accounts, a trademark where your name is used in trade, tight endorsement contracts) makes every later step faster. The legal basis and the orders the Delhi High Court grants are set out on our page on personality rights and deepfake suits.

For advice on your evidence, contracts or a draft complaint, or for updates on the law on deepfakes and the IT Rules, contact Advocate B.K. Singh.