AI Regulation Tracker / National rules

India now requires AI-generated media to carry visible labels and permanent provenance metadata

India's IT Rules amendment on synthetically generated information took effect on February 20, 2026, forcing platforms and publishers to label AI media and embed non-removable provenance identifiers. Any firm putting AI-generated images, audio, or video in front of Indian users is now in scope.

By Anthony Guerriero · Reviewed by The Leveraged Years Editorial Desk · Published July 8, 2026 · Last updated July 8, 2026

India now requires AI-generated media to carry visible labels and permanent provenance metadata regulation briefing

The Leveraged Years AI Regulation Tracker

Regulatory Briefing

Bindingness: Binding Law
Scope: India. Social media platforms, intermediaries, marketers, studios, and platforms, intermediaries, and publishers making AI-generated media available to Indian users.
What Changed: MeitY notified the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, which amend the 2021 IT Rules to regulate "synthetically generated information" (SGI). As reported, the amendments add due diligence duties under Rule 3(1) requiring SGI to be prominently labeled and embedded with permanent provenance metadata that cannot be removed. Significant social media intermediaries must obtain user declarations and verify them by technical means.
Who This Affects: Social media and platform operators, trust-and-safety and compliance teams, marketers and creative studios producing AI images, audio, or video for Indian audiences, and any intermediary enabling upload of such content.
Current Status: In Force
Effective Date: February 20, 2026 (notified February 10, 2026)
Primary Source: MeitY, Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, gazette notification, https://www.meity.gov.in/static/uploads/2026/02/550681ab908f8afb135b0ad42816a1c9.pdf
Secondary Analysis: Chambers and Partners; Mondaq; S.S. Rana & Co.; Khaitan & Co.
Workflow Impact: Teams shipping AI media into India must add labeling and provenance metadata to their production pipeline and, if they run a large platform, build a declaration-and-verification step into upload flows.
Operational Next Step: Map where AI-generated audio, image, or video enters your Indian-facing products and confirm each path applies a visible or audible label plus a permanent provenance identifier.
Last Verified: July 8, 2026

India has moved AI-content labeling from guidance into binding law. On February 10, 2026, the Ministry of Electronics and Information Technology (MeitY) notified the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, which amend the 2021 IT Rules. The amendment took effect on February 20, 2026, and it sets out how "synthetically generated information," or SGI, must be identified when it circulates on Indian platforms. For any company that produces AI media or operates a service where users post it, the compliance question is no longer whether to label AI content, but how to do it in a way the rules will accept.

What counts as synthetically generated information

The amendment defines SGI as audio, visual, or audio-visual information that is created or altered artificially or algorithmically using a computer resource, in a way designed to appear real, authentic, or true and indistinguishable from a real person or a real-world event. As reported by secondary analyses of the notification, the focus is multimodal, covering images, audio, and video rather than text alone. The rules carve out good-faith uses, including text-only content, routine editing, document preparation, and tools that improve accessibility, clarity, translation, or searchability. The target is perceptible harm from realistic fakes, not ordinary AI assistance.

The labeling and provenance duty

The core obligation is specific. SGI must be prominently labeled and embedded with permanent metadata or an appropriate technical provenance mechanism, including a unique identifier. As reported, visual SGI must carry a label that is prominent, easily noticeable, and adequately perceivable, while audio SGI must carry a prominently prefixed audio disclosure. The rules go further than a surface label. They require that the design must not allow the label or the permanent metadata to be modified, suppressed, or removed. In practice that means provenance has to be built into the asset at creation, not bolted on as a caption a downstream user can strip.

Extra duties for large platforms

The amendment places heavier obligations on significant social media intermediaries, a category tied to the registered-user threshold specified under the IT Rules, reported at five million. Under the amended Rule 3(1) due diligence framework, these platforms must obtain a declaration from users on whether uploaded content is synthetically generated, deploy technical means to verify that declaration, and ensure that AI content is clearly displayed and labeled. This turns the platform into an active checkpoint. A user claim that content is or is not AI-generated is not enough on its own, because the platform must also apply its own verification.

What the rules do not do

The amendment does not ban synthetic media, and it does not reach text-only outputs or the good-faith editing and accessibility uses it lists. It also operates through the intermediary framework, so the primary compliance actors are platforms and publishers rather than individual end users. Secondary analyses report short response timelines attached to the wider rules, including takedown windows for unlawful content and for non-consensual or deepfake material, and roughly ten days for intermediaries to implement the technical and operational changes. Firms should confirm the exact timing against the notification text.

Why a US or global firm should read this

The reach is territorial in effect. Any studio, marketer, or platform that publishes AI-generated media to users in India falls within scope, regardless of where the company sits. India is one of the largest internet markets, so its labeling and provenance standard functions as a de facto design requirement for global AI-media pipelines. A firm that builds permanent, non-removable provenance into its Indian output will find it easier to meet similar transparency expectations emerging in other jurisdictions. The pragmatic move is to treat embedded provenance as a default for AI media, not as an India-only feature.

Frequently Asked Questions

What changed under India's IT Rules amendment on synthetic media?

MeitY notified the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026 on February 10, 2026, effective February 20, 2026. AI-generated audio, visual, or audio-visual content, defined as synthetically generated information, must now be prominently labeled and embedded with permanent, non-removable provenance metadata including a unique identifier.

Who is affected by these rules?

Social media platforms and other intermediaries, and platforms, intermediaries, and publishers making AI-generated media available to Indian users, including marketers and creative studios. Significant social media intermediaries, reported as those above five million registered users, carry added duties to collect and technically verify user declarations about AI content.

Does the amendment apply to AI-generated text?

As reported, no. The definition centers on audio, visual, and audio-visual content designed to appear real and indistinguishable from a real person or event. The rules exclude text-only content and good-faith uses such as routine editing, document preparation, translation, and accessibility improvements.

What must a platform do with user-uploaded content?

Under the amended Rule 3(1) due diligence framework, a significant social media intermediary must obtain a user declaration on whether uploaded content is synthetically generated, deploy technical means to verify that declaration, and ensure AI content is clearly displayed and labeled rather than relying on the user's claim alone.

How should a global company operating in India respond now?

Map every path where AI-generated audio, image, or video reaches Indian users and confirm each applies a visible or audible label and a permanent provenance identifier that cannot be stripped. Verify exact timelines and thresholds against the MeitY notification, since the gazette text is the controlling source.