Do AI-generated ads and images need a label in India?

What do ASCI's AI guidelines say?

ASCI, the Advertising Standards Council of India, defines synthetically generated content as audio, visual or audio-visual material made or altered by AI so that it appears real and shows a person or event in a way likely to be believed as true. That covers images as well as video and sound¹. The guidelines sort uses into three groups.

First, some uses are prohibited even with a label. The guidelines name fabricated endorsements or testimonials, AI visuals that exaggerate product results, non-existent locations shown as real, and unauthorised copyrighted work, deepfakes or a person's likeness without consent¹.

Second, a label is mandatory when the AI content materially influences a consumer's decision and leaving the label out would mislead. ASCI's examples are below.

Third, no label is needed where AI has no material effect on an informed choice, as listed in the facts above. Where a label is needed, brands may use their own wording or the platform's, such as "Audio/Video created using AI"¹. A label does not cure a misleading ad: the whole ASCI Code still applies¹.

• Synthetic influencers and ambassadors.
• Replicating a real person's likeness or voice, even with their consent, for personalised messaging.
• Fabricated events, settings or situations that affect how the product or service is understood.
• Demonstrating a product that does not currently exist, such as a model of an unbuilt housing complex.
• AI sound effects that are highly relevant to the product's core features.
• Paid or sponsored AI product suggestions, labelled as sponsored by the brand.

When do the ASCI rules start?

The date is unconfirmed. The ASCI document says the guidelines take effect three months after publication but states no calendar date, and the PDF is signed 17 September 2026¹. Trade press reported the release on 29 September 2026⁴. Counting three months from those two dates gives mid to late December 2026. Outlets may report a different day, so plan to be ready by the end of December 2026 and check ASCI's own notice for the exact date.

What do India's IT Rules 2026 require, and from whom?

The February 2026 amendment defines synthetically generated information as audio, visual or audio-visual information made or altered by computer so that it appears real and is indistinguishable from a natural person or real-world event. It excludes routine editing, colour adjustment, noise reduction, compression, and uses that only improve accessibility or translation².

The duties sit on intermediaries. Under Rule 3(3), a platform that offers a tool for creating synthetic content must use technical measures to block unlawful material, label the rest prominently, embed permanent metadata where feasible, and not let the label be removed. Under Rule 4(1A), a significant social media intermediary must require users to declare whether content is synthetic, verify the declaration, and label confirmed content².

The rule text names intermediaries, not advertisers. In practice, expect a declaration prompt when you upload realistic AI media to a large platform, and answer it honestly².

One change is still only a draft. A MeitY draft placed in the public domain on 30 March 2026 would replace the current prominent-visibility wording with a label that stays visible throughout a video, with comments open until 7 May 2026. MediaNama reported it as a draft, and we found no final notification, so its status is unconfirmed⁵.

Who carries the duty: the brand, the agency or the platform?

ASCI's chief executive said responsibility stays with advertisers to make sure the final communication is honest, transparent and compliant with the ASCI Code⁴. The IT Rules, by contrast, bind platforms and tool providers². An ad can fall under both.

If an agency or studio makes the ad, agree in writing who decides whether a label is needed, who writes it, and who keeps the record of which parts were AI-generated.

What about the US and the EU?

In the EU, Article 50 of the AI Act requires deployers of an AI system that generates or alters image, audio or video resembling real people, objects or places to disclose that it is artificially generated. Evidently artistic, creative or fictional works get a lighter duty. The article's page lists 2 August 2026 as its date⁶. This page does not cover the separate duties on AI providers to mark their output.

In the US, the federal item we cite is the FTC's final rule of 14 August 2024, which bans fake and AI-generated reviews and testimonials that misrepresent a person who does not exist⁷. New York requires ads using a synthetic performer to disclose it conspicuously from 9 June 2026, with civil penalties of $1,000 for a first violation and $5,000 for later ones⁸. Google Ads names the EU, India and New York as places with AI disclosure rules⁹.

What should I do before I publish?

Work through each AI element of the ad, not the ad as a whole.

• List the AI elements: people, voices, scenes, products, music, sound effects.
• For each, ask whether it changes what a buyer would believe about the product. If yes, label it. If it fits ASCI's no-label list, you can leave it.
• Do not use AI to make a testimonial, a before-and-after result, or a place that does not exist. A label does not make these acceptable¹.
• Keep written consent for any real person whose face or voice is replicated.
• Burn the label into the ad and use each platform's own AI label setting where one exists. Google says its label setting does not guarantee compliance with specific regulations⁹.
• Have a lawyer review anything high-risk, such as health, finance or endorsements.