AI Likeness Consent for Ads in India After Delhi HC Order
On 6 November 2026, the Delhi High Court ordered the takedown of six AI-generated videos and posts that showed a real person in fabricated, demeaning scenes. The order was reported on 14 November 2026. For anyone who makes AI ads, avatars or promo videos using real faces and voices, AI likeness consent is now a practical question rather than a theory. This guide explains what the court said, what Indian rules say about synthetic media in ads, and how to build a simple consent check before you publish.
What the Delhi High Court decided
The case is Rekha Gupta v. Aam Aadmi Party (CS(OS) 1019/2026), heard by Justice Manoj Jain. The suit targeted AI-generated videos, images and posters published from the official and verified accounts of the party. They portrayed the Chief Minister of Delhi as a liquor vendor, a hawker and a delivery agent, and as the head of a non-existent scheme. One video was set to a romantic soundtrack and showed her serving liquor at night during a brawl. According to the court’s reporting, the court directed the takedown of six items: five posts or videos on X and one video on Facebook.
The court was clear that political criticism and satire are part of free speech, and that public office-holders must tolerate them. It also said there is “a thin line between political satire and offensive depiction”. Satire, it held, cannot be used to reduce a person to a figure of mockery. Because the videos looked near-realistic, the court found that they threatened personal dignity. It noted the lack of stringent legislation on AI content and said courts must scrutinise AI-generated political material when dignity-based rights are alleged to be infringed.
This was an interim finding, not a final judgment. The matter is listed for 4 December 2026. The point for advertisers is narrower but important: a realistic AI clip of a real person, shown doing something they never did, can be taken down quickly, even when the maker says it is only a parody.

Why AI likeness consent matters under Indian rules
Indian rules treat synthetic media differently from ordinary editing. The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026 came into force on 20 February 2026. They define synthetically generated information (SGI) as audio, visual or audio-visual material that a computer resource creates or alters so that it appears real and could be mistaken for a real person or event. Deepfakes, AI-altered photos and videos, and voice clones are covered. AI-generated text is not SGI under these rules, although existing duties on unlawful text still apply.
Non-prohibited SGI must be clearly and prominently labelled. Visual content needs a visual label and audio needs an audio disclosure. Labels and metadata must not be removed or altered. The rules also shortened takedown windows. For lawful court or government orders about unlawful content, platforms must act within roughly three hours, down from 36 hours. For high-risk categories such as impersonation and morphed images, the window is two hours, down from 24.
The advertising regulator’s guidance is more specific to ads, but it is still a draft. A May 2026 report on the Advertising Standards Council of India’s draft guidelines says a real person’s likeness or voice used for personalised messaging must be labelled, even with consent. Synthetic visuals that show product performance must also be labelled unless they accurately reflect actual performance. The same draft lists several things as prohibited regardless of any label, including fake endorsements or testimonials, deepfakes or unconsented use of someone’s likeness, and fictional authority figures such as a fake doctor. The report gives no effective date, so treat the guidance as a signal of direction rather than settled law. It also says that a correct label does not make misleading or harmful content acceptable.
Put simply, the law asks two separate questions. The first is whether the content is labelled. The second is whether the person shown agreed to be shown that way. A label answers the first question and says nothing about the second.
A consent check before any AI ad goes live
Use this short checklist for every AI-made creative that shows a real face, voice or identifiable person:
- Get written permission from the person whose face or voice appears, and state exactly what the clip will be used for, where, and for how long.
- Never show a real customer, doctor, influencer or public figure endorsing a product unless they have agreed in writing to that specific use.
- Do not show a product doing something it cannot do, or a result that the product has not produced.
- Label the creative visibly as AI-made, even if your tool already adds a label, because a platform’s detection can miss an edit.
- Keep the prompt, the source photos, the consent note and the approval record for each ad, so you can answer a complaint quickly.
- Check the disclosure settings of the marketplace or ad platform before you launch.
- For health, finance or food claims, have a lawyer review both the words and the visuals.
A consent record is only as good as the process behind it. The simplest rule is to treat every real person in an AI clip as someone who must say yes first, and to keep a copy of that yes.

How InstaStudio helps, and where it stops
InstaStudio is an India-first AI creative studio for creators, sellers and small businesses. Its face-swap tool is listed in the product’s feature set as requiring consent, and the product does not replace your own consent process. The studio also has several controls that matter when a real person is involved. Outputs carry a visible AI label on paid plans, while the free plan adds a watermark. Outputs also include IPTC AI metadata and signed C2PA Content Credentials. The trust-list certificate that makes C2PA credentials verifiable is still pending, so do not treat those credentials as proof of origin yet.
For complaints, InstaStudio has a /report grievance and takedown form with 2-hour, 3-hour and 24-hour deadlines, feeding an admin queue. It also keeps DPDP consent records and supports data export and account deletion. These features help you respond when something goes wrong. They do not decide whether a person agreed to appear in your ad.
Pricing is simple, and every price below includes GST. A Day Pass gives Pro features for 24 hours for ₹199. A 500-credit pack costs ₹79. Basic costs ₹799 per month with 5,500 credits, removes the watermark and includes commercial use. You can compare plans on the InstaStudio product page.
What to watch before 4 December
Three things are worth tracking. The first is the next hearing in the Delhi case on 4 December 2026, which may show how the court treats AI content in its final decision. The second is whether the advertising regulator confirms its final guidance on AI disclosures in ads, since the May 2026 report describes a draft. The third is whether MeitY notifies any change to its labelling rules. Until those land, the safest approach is the same one the current rules already point to: visible labels, clear consent and a fast route to takedown.
Frequently Asked Questions
Does the Delhi High Court order apply to AI ads?
The order concerned political videos and posters, and it was an interim finding. It does not set a general rule for product ads. It does show that realistic AI content about a real person can be taken down quickly, and that courts will look at dignity and realism, not only at whether a label was added.
Is a visible AI label enough if I have the person’s permission?
Not on its own. A label tells viewers the content is AI-made, but it does not make a fake endorsement or an unconsented use of someone’s likeness acceptable. The draft ASCI guidance lists those as prohibited regardless of any label.
Do I need written consent to use an AI version of a real person?
Written consent is the safest standard and the one we recommend for every real face or voice in an ad. Keep a copy that states the use, the channels and the duration. Check your own lawyer’s advice for your situation, especially for regulated categories.
Does InstaStudio check consent for me?
No. The product provides labels, metadata, a takedown form and consent records for your account, but responsibility for getting permission from the person shown stays with you. Use the checklist above before every shoot or generation.
Sources
- SCC Online: Delhi High Court orders takedown of AI-generated deepfakes of CM Rekha Gupta (14 November 2026)
- Freshfields: India targets deepfakes and AI-generated content, key changes under MeitY’s 2026 amendments to the IT Rules
- Legal500: IT Rules 2026 deepfake regulation, three-hour takedowns and AI labelling obligations
- Indian Television: India’s advertising watchdog sets new rules on AI-generated content in ads (May 2026 report on ASCI draft guidelines)
Want AI creatives with visible labels, consent records and a takedown route built in? Try InstaStudio, or read more guides on the InstaDataHelp blog.
