Indian advertisers now face a three-month clock. The Advertising Standards Council of India (ASCI) published guidelines on 29 September that require a label on synthetic images, video and audio when the content “materially influences consumer decisions and the absence of disclosure could mislead consumers.”
The test is narrower than “all AI content.” A synthetic background or routine colour correction generally does not need a label; a synthetic spokesperson, a consent-based celebrity voice clone, or a fabricated product demo can.
What brands must label
- Synthetically generated influencers and ambassadors.
- Replication of a real person’s likeness or voice with consent for personalised messaging.
- Fabricated events or settings that could change how consumers see the advertised product or service.
- Demonstrations of products that do not currently exist.
- AI-generated sound effects central to a product’s core features, such as headphones.
- Paid or sponsored AI-generated product suggestions, labelled “Sponsored by [Brand]”.
Even a label does not make everything acceptable. Fabricated endorsements or testimonials, misleading synthetic exaggeration, fake locations presented as real, unauthorised copyrighted work, and deepfakes using a person’s likeness without consent remain banned.
Routine edits, ambient or decorative elements, obvious fantasy effects, ad copy generation, and accessibility uses such as accurate subtitles and audio descriptions are exempt.
Why platform labels are not enough
Since 20 February, the IT Amendment Rules 2026 have required platforms to label AI-generated content. That is a platform-level duty: large platforms with over 50 lakh users in India must ask uploaders to declare AI content, verify it and label it. ASCI’s rules instead put the duty on the advertiser across TV, print, outdoor and online.
The triggers differ too. IT Rules ask whether content is AI-generated; ASCI asks whether leaving it unlabelled could mislead a buyer. A platform label may appear on an ad even when ASCI does not require a brand label, and ASCI has not confirmed that a platform’s own label meets the advertiser’s responsibility.
“Importantly, it clarifies where the mere use of an AI label may not make an otherwise misleading advertisement acceptable,” said Manisha Kapoor, ASCI Secretary General and CEO.
Ambiguities to watch
A promoted AI music video is both product and ad. Saregama has said it uses AI to create fresh videos for older catalogue at costs as low as Rs 70,000 per video. Whether a paid promotion of such a video needs an ASCI label depends on what it shows, whether it recreates a real person, and whether the synthetic content could mislead a buyer about the video itself.
An ad stitched from multiple AI clips raises a similar puzzle. A 26-second ad may be assembled from six to twelve clips of two to four seconds each; fixes can cost roughly Rs 4,000–10,000 per video. ASCI does not say whether brands should label the full ad or only the synthetic portions.
AI chatbots add another gap: OpenAI’s ChatGPT ads in India carry a paid label, but ASCI does not spell out how an advertiser should label a product suggestion generated by the AI system itself.
What ad teams should do
ASCI’s decisions carry no legal penalty, but many banned uses can already be enforced through CCPA guidelines, copyright law and IT Rules deepfake provisions. Run a three-part audit before a campaign ships: does the creative show a real-looking person or event, does it imply endorsement or proof, and is it central to the product claim? If yes, label conservatively and document consent, especially for voice clones.
Also watch MeitY. Its labelling requirements have shifted from a 10% screen coverage draft to a “prominent” label under February 2026 rules, and an April 2026 proposal may require labels to stay continuous and visible for the full length of a video.
Source: MEDIANAMA
