In short
- ASCI's guideline (signed 17 Sep 2026, effective three months after publication, so around mid-December) requires a label where AI content materially influences consumer decisions and its absence would mislead.
- Synthetic influencers need a label. Fabricated testimonials, deepfakes, likeness without consent and AI-exaggerated results are prohibited even with a label.
- The IT Rules amendment (G.S.R. 120(E), in force 20 Feb 2026) puts labelling duties on platforms. Significant social media platforms must ask users to declare synthetic content before it is published.
- An AI label sits beside the ad label, not instead of it. A synthetic spokesperson in a paid post carries both, plus the virtual-influencer disclosure.
For influencer content, India has two AI rules that matter. ASCI's Guidelines for Responsible Labelling of Synthetically Generated Content in Advertising, signed on 17 September 2026, require a label where AI content materially influences consumer decisions and leaving it out would mislead. The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules 2026, notified on 10 February 2026, put labelling and declaration duties on platforms.
This post explains what each rule asks of a brand and a creator, with scenarios. It sits below the AI section of our ASCI influencer guidelines guide, and the wider process is in our complete guide to influencer marketing in India.
Two rules, two audiences
| ASCI guideline | IT Rules amendment | MeitY draft | |
|---|---|---|---|
| Aimed at | Advertisers and ads, including influencer ads | Platforms (intermediaries) | Platforms |
| Date | Signed 17 Sep 2026; effective three months after publication | Notified 10 Feb 2026 as G.S.R. 120(E); in force 20 Feb 2026 | Draft dated 30 Mar 2026 |
| Force | Self-regulatory code | Law | Proposal |
| Asks | Label AI content that materially influences decisions; prohibits certain uses | Prevent unlawful synthetic content, label the rest, ask for user declarations | Continuous display of labels |
| Our confidence | High on content; start date is arithmetic | High on text and date | Low; we did not find it notified |
PIB's release of 25 March 2026, giving information from MeitY, confirms the amendments were notified on 10 February 2026 and came into force on 20 February 2026. The same release says they mandate clear labelling and traceability of permissible synthetic content. The MeitY draft of 30 March 2026 proposes continuous display of synthetic content labels. Consultation ran to 29 April 2026. We did not find that it has been notified.
What ASCI counts as synthetic content
ASCI defines synthetically generated content, or SGC, as audio, visual or audio-visual content in an advertisement that is artificially or algorithmically created, generated, modified or materially altered using a computer resource, to appear real, authentic or true, and that depicts a person or event in a way likely to be believed true or to be indistinguishable from a natural person or real-world event.
To decide whether a label is needed, ASCI gives three questions:
- Is the SGC central to the ad's persuasive messaging, or does it materially influence purchase decisions?
- Would the way it is used mislead consumers about the product's claims, performance or benefits if it were not labelled?
- Would leaving out a disclosure create a false or misleading impression of claims, benefits or performance?
ASCI says its focus is consumer outcomes, not the technology, and that every ad containing SGC must still comply with the whole ASCI Code.
The three buckets
| Needs a label | Prohibited even with a label | No label needed |
|---|---|---|
| Synthetically generated influencers and ambassadors | Fabricated endorsements or testimonials | Minor enhancements: colour correction, light adjustment, noise reduction, standard blemish removal, formatting, copy refinement, if they do not change the ad's substance or core claims |
| A real person's likeness or voice replicated, even with consent, in ASCI's personalised-messaging example | AI visuals that exaggerate results, such as an AI before-and-after weight-loss image | Decorative backgrounds, ambient music, jingles and background sound effects unrelated to product capabilities |
| Fabricated events or settings that affect understanding of performance | Non-existent locations shown as real | Obviously fantastical effects |
| A product demonstration for a product that does not yet exist | Unauthorised copyrighted work, deepfakes, or a person's likeness without consent | Generating or enhancing ad copy, audio descriptions or documents in good faith |
| Sound effects highly relevant to core product features | Accessibility uses: subtitles, captions, translations, accurate audio descriptions | |
| AI-generated paid or sponsored product suggestions, labelled "Sponsored by [Brand]" |
Creator scenarios
These are our readings of ASCI's examples. They are not rulings.
| Scenario | Our reading |
|---|---|
| A creator uses AI to draft a caption or script | No label. This is an administrative and text use |
| AI-generated subtitles in Hindi, Tamil or Telugu | No label. Accessibility use |
| Colour grading, noise removal or background clean-up | No label, as long as the product's look is not changed |
| A skin-smoothing filter on a skincare "results" Reel | Risky. Blemish removal is a minor enhancement, but if it changes how results look it exaggerates them, which ASCI treats as prohibited even with a label |
| An AI avatar reads a brand script as a spokesperson | Label needed, as a synthetic influencer or ambassador, plus the virtual-influencer disclosure |
| An AI avatar says "I used this for 30 days" | A fabricated testimonial. Prohibited even with a label |
| A creator's cloned voice dubs an ad into another language | Unclear. Label it and get written consent |
| An AI render of a product that is not built yet | Label needed |
| An AI image of a hotel room or location that does not exist, shown as real | Prohibited |
| A paid recommendation delivered by an AI assistant | Label as "Sponsored by [Brand]" |
The skin-smoothing row is the one we see most often in beauty content. Our page on beauty and skincare influencer marketing covers claims in that category, and our post on health influencer rules covers the weight-loss example.
Label wording, and the second label
ASCI says brands may use labels provided by ad platforms or their own, such as "Audio/Video created using AI" or "Audio/Video enhanced using AI", or alternatives that accurately inform the consumer. For paid AI recommendations the label must say "Sponsored by [Brand]". Disclaimers should follow ASCI's disclaimer guidelines where relevant.
An AI label never replaces the ad label. A creator post that is paid for needs the influencer-advertising label under ASCI's other guideline, upfront and for the required time, as our post on disclosure labels sets out. If the AI content is a virtual influencer, ASCI also requires an upfront, prominent statement that the viewer is not interacting with a real human. That is three possible disclosures on one asset. Plan the layout so they do not stack on top of each other. Our post on virtual influencers in India covers the third.
What the IT Rules amendment actually says
The MeitY consolidated text, updated to 10 February 2026, adds a definition of "synthetically generated information": audio, visual or audio-visual information artificially or algorithmically created, generated, modified or altered using a computer resource so that it appears real, authentic or true and depicts a person or event in a way that is, or is likely to be perceived as, indistinguishable from a natural person or real-world event. It excludes routine or good-faith editing, formatting, enhancement, colour adjustment, noise reduction, transcription or compression that does not materially alter the substance, and it excludes use solely for accessibility, clarity, translation or discoverability without generating, altering or manipulating any material part of the underlying content.
It then puts two sets of duties on platforms:
- Platforms that offer AI creation tools must use reasonable technical measures to stop unlawful synthetic content. That includes content that falsely depicts a real person's identity, voice, conduct or statements in a way likely to deceive. They must label all other synthetic information prominently, embed permanent metadata or a provenance mechanism with a unique identifier where technically feasible, and not let users remove the label.
- Significant social media intermediaries must, before display or upload, require users to declare whether the content is synthetic, use technical measures to verify the declaration, and label content that is confirmed as synthetic.
These duties sit on platforms and not on creators. The practical effect is a declaration prompt when a creator uploads, and a platform label that appears on the post whether or not the brand wants it. Which platforms count as "significant" depends on a threshold the government notifies, which we did not check.
What brands should do now
- Ask for AI declarations. In every brief response, ask creators to list any AI tool used on visuals, voice or likeness.
- Write it into the contract. Require disclosure of AI use, bar synthetic versions of the creator or any real person without separate written consent, and warrant that any testimonial reflects real use. Our guide to influencer contracts in India lists the clauses.
- Cover likeness separately. Ordinary usage rights do not cover new content made from a creator's face or voice. See usage rights for influencer content.
- Keep originals. Store the source footage next to the edited file. If a complaint arrives, you can show what was changed.
- Spec the labels. Put the exact AI label, ad label and their positions in the brief.
- Label your own AI. A synthetic spokesperson or AI-generated recommendation made by the brand needs a label too.
- Never generate reviews. AI reviews and AI-made "customer" clips are the clearest prohibited use.
- Put mid-December 2026 in the calendar. Check ASCI's site for the exact effective date, and brief creators before it.
If you run AI-dubbed regional content, see regional influencer marketing, where this question will come up first.
Dates and confidence
| Item | Date | Status | Confidence |
|---|---|---|---|
| ASCI SGC guideline | Signed 17 Sep 2026 | Effective three months after publication, about mid-December 2026 by our count | High on content; date is arithmetic |
| IT Rules amendment, G.S.R. 120(E) | Notified 10 Feb 2026; in force 20 Feb 2026 | In force | High |
| MeitY draft on continuous label display | 30 Mar 2026 | Proposal; not found notified | Low |
| AI dubbing and voice cloning in creator ads | n/a | Unresolved | Low |
To plan creator content with AI rules built into the brief, see how we work or send us a brief.
This is general information, not legal advice.
Sources
- The ASCI Code and guidelines page: Guidelines for Responsible Labelling of Synthetically Generated Content in Advertising (signed 17 Sep 2026) and Guidelines for influencer advertising in digital media, ASCI (undated page, accessed 31 Oct 2026)
- Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, updated as on 10.02.2026, with insertions by G.S.R. 120(E), MeitY (10 Feb 2026)
- Government committed to fostering AI innovation and growth of digital platforms in a responsible and secure manner (reply in Rajya Sabha, with information from MeitY), PIB, Ministry of Women and Child Development (25 Mar 2026)
- Draft amendment on continuous display of synthetic content labels, MeitY (30 Mar 2026)
Frequently asked questions
Do influencers have to label AI-generated content in India?
In advertising, yes, where the AI content materially influences consumer decisions and leaving out a label would mislead. ASCI's guideline names synthetic influencers and replicated real voices as examples. Routine edits, decorative backgrounds, obvious fantasy and accessibility uses need no label. Platforms also have their own duties under the IT Rules. See our ASCI guide.
When does ASCI's synthetic content guideline take effect?
ASCI's note says it takes effect on the expiry of three months from the date of publication. The guideline is dated 17 September 2026, so counting from that date gives about mid-December 2026. That is our arithmetic, so confirm the exact date on ASCI's website before relying on it.
What label wording can a brand use for AI content?
ASCI says brands may use platform-provided labels or their own, such as 'Audio/Video created using AI' or 'Audio/Video enhanced using AI', or any wording that accurately informs the consumer. For an AI-generated recommendation that is paid for, the label must read 'Sponsored by [Brand]'. The ad label for the influencer post is still needed.
Is AI dubbing or AI translation exempt from labelling?
Unclear. ASCI exempts accessibility uses such as subtitles and translations, and the IT Rules definition excludes translation done solely for accessibility without generating or altering material content. A cloned voice speaking words the creator never recorded goes further. Until ASCI clarifies, label it and get the creator's written consent.
Can a brand use an AI-generated customer review or testimonial?
No. ASCI lists fabricating endorsements or testimonials, using AI to create or attribute an endorsement that falsely represents a real person as having used or recommended a product, as prohibited even with an AI label. An AI avatar that says it used your product for 30 days has the same problem. See UGC content creators.
Do platforms label AI content automatically in India?
The IT Rules require platforms that offer AI creation tools to label synthetic output and embed provenance data where technically feasible. Significant social media intermediaries must ask users to declare synthetic content before publication and label confirmed content. Expect a declaration prompt at upload and answer it truthfully. Which platforms count as significant depends on a notified threshold.