Published 16 September 2026
Since 20 February 2026, anyone uploading an AI-generated image, video or audio clip to a large social media platform in India has been required to declare it, and the platform has been required to label it. The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, notified by the Ministry of Electronics and Information Technology as G.S.R. 120(E) on 10 February 2026, add a defined category, "synthetically generated information", to the 2021 Rules and attach labelling, metadata and takedown duties to it. The rules were written for deepfakes, but they catch the AI product shots, generated voice-overs and synthetic presenters that Indian businesses now use routinely in marketing. This post explains what the rules require, where the obligation falls, and what a brand that uses AI in its content should change.
What counts as synthetically generated information
New Rule 2(1)(wa) defines synthetically generated information as audio, visual or audio-visual content that is artificially or algorithmically created, generated, modified or altered in a way that appears authentic or true, including where it depicts a person or event. Routine editing is carved out: colour correction, compression, cropping, noise reduction and similar good-faith adjustments that do not change what the content shows are not synthetic; nor is watermarking for branding, accessibility processing, or academic and research material that does not impersonate. The line runs between adjusting a real recording and generating or altering one so that it shows something that did not happen. An AI-rendered image of a product on a beach the product has never visited is on the synthetic side of that line; a retouched photograph of the product is not.
The three obligations
- Label and metadata (Rule 3(3)): an intermediary that enables the creation or sharing of synthetic content must ensure it carries a prominent label or notice, embed permanent metadata or a unique identifier recording its synthetic origin and any alteration, and must not allow the label or metadata to be removed or suppressed. A visual label must be clearly displayed; an audio label must be audible in the opening portion.
- Declaration and verification (Rule 4(1A)): a significant social media intermediary, a platform with over fifty lakh registered Indian users, must ask every user at upload whether the content is synthetically generated, verify the declaration with automated tools, and display the label where the declaration or the check says yes. A user who declares falsely is exposed under the platform's terms and the general criminal law.
- Faster takedown (Rule 3(1)(d)): content flagged by a court order or a reasoned government intimation must be removed within three hours instead of thirty-six, and grievance complaints must be resolved within seven days instead of fifteen. A platform that does not comply loses the safe-harbour protection of section 79 of the IT Act for that content.
Where this lands on a brand
The rules bind intermediaries, and through the declaration requirement they bind users who upload, which includes the marketing manager or agency posting on a company's behalf. There is no separate offence for a business that publishes an unlabelled AI image on its own website; the obligation bites when the content is uploaded to Instagram, YouTube, Facebook, X, LinkedIn or a similar platform, or placed as an ad through their systems, and the declaration is made or not made. Two consequences follow. First, an AI-generated product film or synthetic presenter posted to a platform will carry a visible synthetic-content label from the platform whether the brand likes it or not, which changes how the creative reads to the audience. Second, a false declaration, or stripping metadata before upload, is the kind of conduct that costs an account and invites a complaint under the Bharatiya Nyaya Sanhita where the content deceives. The Advertising Standards Council of India's existing rule that advertisements must not mislead applies to AI creative exactly as to any other.
A working policy for AI in marketing
- Keep a register of which assets are AI-generated or AI-altered, tagged at creation, so the person uploading knows what to declare.
- Declare honestly at upload on every platform, and instruct agencies in writing that declarations are made on the brand's behalf and metadata is never stripped.
- Never generate a real person's likeness or voice, a customer's, an employee's or a public figure's, without written consent; the rules' impersonation provisions and the ordinary law on personality rights both apply.
- For testimonials, before-and-after images and product demonstrations, use real footage; a synthetic label on a testimonial reads as a confession.
- Add a short disclosure line to AI-generated visuals on the brand's own website and emails even though the rules do not require it there, since the same asset will be labelled on the platforms and inconsistency looks like concealment.
Frequently Asked Questions
When did the AI content labelling rules take effect in India?
On 20 February 2026. The IT (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026 were notified as G.S.R. 120(E) on 10 February 2026.
Does a business have to label AI images on its own website?
The rules bind intermediaries and users who upload to them, not a business's own site. The label and declaration duties apply when the content is posted to a significant social media platform or run as an ad through one.
Is a retouched product photo synthetically generated information?
No. Routine editing such as colour correction, cropping, compression and noise reduction that does not change what the content depicts is excluded from the definition; a generated scene or altered event is included.
What is a significant social media intermediary?
A social media platform with more than fifty lakh registered users in India, the threshold notified under the 2021 Rules. The declaration-and-verification duty in Rule 4(1A) applies to these platforms.
What happens if a platform fails to label or remove synthetic content?
It risks losing the safe-harbour immunity of section 79 of the IT Act for that content, which exposes it to liability for what the user posted.
Sources
Ministry of Electronics and Information Technology, Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, G.S.R. 120(E) dated 10 February 2026, in force 20 February 2026, amending Rules 2, 3 and 4 of the 2021 Rules; commentaries by Chambers and Partners, RNA Technology and IP Attorneys and Mondaq cross-checked for rule numbers and timelines; Information Technology Act, 2000, section 79. Read the notified text before relying on a summary.
BookMyTM advises brands in Kerala on trademark, copyright and advertising compliance, including the use of AI-generated content in marketing.