Who Must Label AI-Generated Content in India? Understanding Intermediary Obligations Under the IT Rules, 2026 

Posted On - 17 November, 2026 • By - Dhruv Kaushal

Introduction

The increasing use of artificial intelligence (“AI”) to create realistic images, videos and audio recordings has raised important questions about the responsibility for identifying and disclosing synthetic content. In India, the regulatory response has emerged through amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 (“IT Rules”), rather than through a standalone law governing artificial intelligence. 

The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, notified by the Ministry of Electronics and Information Technology (“MeitY”) on 10 February 2026, introduced a dedicated framework governing synthetically generated information (“SGI”). Effective from 20 February 2026, the amendments impose due diligence obligations on intermediaries that facilitate the creation, modification, publication or dissemination of synthetic media. They also prescribe enhanced verification and labelling requirements for Significant Social Media Intermediaries (“SSMIs”). 

The framework raises an important compliance question: does the responsibility to label AI-generated content rest with the individual who creates it, the platform that hosts it, the intermediary that provides the AI-generation tool, or all of them in different capacities? 

The answer depends on the nature of the service being provided, the role of the intermediary and the type of content involved. Understanding this allocation of responsibility is essential for AI tool providers, social media platforms, digital publishers and businesses that use or commission synthetic media. 

What qualifies as synthetically generated information?

Rule 2(1)(wa) of the IT Rules defines “synthetically generated information” as audio, visual or audio-visual information that is artificially or algorithmically created, generated, modified or altered using a computer resource, in a manner that makes it appear real, authentic or true and depicts or portrays an individual or event in a way that is, or is likely to be perceived as, indistinguishable from a natural person or real-world event. 

The definition is therefore not limited to content generated entirely through artificial intelligence. It also extends to existing audio, visual or audio-visual material that has been modified or altered using a computer resource, provided the resulting content satisfies the prescribed authenticity and indistinguishability threshold. This distinction is important because not every use of AI or digital editing automatically qualifies as SGI. The definition is directed towards realistic synthetic representations of individuals or events, rather than every form of computer-assisted content creation. 

Express exclusions from the definition

The Rules provide specific exclusions to prevent routine and good-faith uses of technology from being treated as SGI. These include: 

  • Routine editing and enhancement: Editing, formatting, technical correction, colour adjustment, noise reduction, transcription or compression that does not materially alter, distort or misrepresent the substance, context or meaning of the underlying content. 
  • Preparation of documents and educational material: Routine or good-faith creation, preparation, formatting or design of documents, presentations, PDFs, educational or training materials and research outputs, including illustrative, hypothetical, draft or template-based content, provided the process does not generate a false document or false electronic record. 
  • Accessibility and quality improvements: Use of computer resources solely to improve accessibility, clarity, quality, translation, description, searchability or discoverability, without generating, altering or manipulating any material part of the underlying audio, visual or audio-visual information. 

For example, improving the brightness of a photograph, removing background noise from an interview, adding subtitles to a video or translating a speech without materially changing its content would ordinarily fall within the exclusions. By contrast, using AI to create a realistic video falsely depicting an individual making a statement they never made may qualify as SGI. The assessment must therefore focus on the nature and effect of the alteration, rather than merely whether an AI tool was used. 

Who bears the labelling obligation?

The 2026 amendments place the principal SGI-related due diligence obligations on intermediaries. However, the scope of the obligation is broader than social media platforms alone. 

Intermediaries facilitating the creation or dissemination of synthetic media

Rule 3(3) applies where an intermediary offers a computer resource that may enable, permit or facilitate the creation, generation, modification, alteration, publication, transmission, sharing or dissemination of SGI. 

This formulation is significant because it covers intermediaries at different stages of the synthetic-content lifecycle. Depending on the nature of the service, it may include: 

  • Platforms providing AI-based image and video generation tools; 
  • Services enabling voice synthesis or voice cloning; 
  • Platforms facilitating the manipulation of existing audio-visual material; 
  • Social media platforms hosting or disseminating synthetic content; and 
  • Other intermediary services that facilitate the publication or distribution of SGI. 

The obligation is therefore not confined to the platform on which a user ultimately uploads the content. An intermediary providing the computer resource used to create or modify synthetic media may also fall within the scope of Rule 3(3). 

Such intermediaries must deploy reasonable and appropriate technical measures, including automated tools or other suitable mechanisms, to prevent users from creating, generating, modifying, publishing, transmitting, sharing or disseminating SGI that violates applicable law. For SGI that is not covered by the prohibited categories under Rule 3(3)(a)(i), the intermediary must ensure prominent labelling and appropriate technical provenance mechanisms. 

Additional obligations for Significant Social Media Intermediaries

SSMIs are subject to an additional layer of compliance under Rule 4(1A). Before allowing information to be displayed, uploaded or published on their computer resources, an SSMI must: 

  1. Require users to declare whether the information is synthetically generated; 
  1. Deploy appropriate technical measures, including automated tools or other suitable mechanisms, to verify the accuracy of the declaration, having regard to the nature, format and source of the information; and 
  1. Where the declaration or technical verification confirms that the information is SGI, ensure that it is clearly and prominently displayed with an appropriate label or notice. 

The Rules further clarify that an SSMI’s responsibility extends to taking reasonable and proportionate technical measures to verify user declarations and ensure that SGI is not published without the required declaration or label. This creates a more demanding compliance standard for SSMIs than the general due diligence obligations applicable under Rule 3(3). An SSMI cannot rely exclusively on a user’s declaration without implementing appropriate verification measures. 

What responsibilities do content creators and uploaders have?

Although the specific labelling and technical provenance obligations under Rule 3(3) are imposed on intermediaries, the Rules do not leave content creators and uploaders without responsibilities. Under Rule 3(1)(b), intermediaries are required to inform users that they must not host, display, upload, modify, publish, transmit, store, update or share information falling within the prohibited categories specified under the Rules. These restrictions include information that is unlawful, harmful to children, invasive of privacy, impersonates another person or deceives or misleads users. 

Rule 3(1)(c) also requires intermediaries to periodically inform users, at least once every three months, about their obligations under the platform’s rules and the potential consequences of non-compliance. Further, Rule 3(1)(ca) requires intermediaries facilitating the creation, generation, modification, alteration, publication, transmission, sharing or dissemination of SGI to provide additional warnings about the legal consequences of using those resources to create or disseminate unlawful synthetic content. 

Where an intermediary becomes aware of a violation relating to SGI covered under Rule 3(3)(a)(ii), Rule 3(1)(cb) requires it to take expeditious and appropriate action in accordance with the Rules. 

Accordingly, the regulatory framework distinguishes between the intermediary’s obligation to implement labelling and technical safeguards and the user’s responsibility not to create or disseminate unlawful content. The fact that an individual is not directly assigned the Rule 3(3) labelling obligation does not mean that the individual is free from other applicable legal or platform obligations. 

How must permitted synthetic content be labelled?

The February 2026 amendments distinguish between synthetic content that falls within the prohibited categories and SGI that may lawfully be created or disseminated. For SGI not covered by the prohibited categories under Rule 3(3)(a)(i), Rule 3(3)(a)(ii) requires the intermediary to ensure that the content is clearly and prominently labelled. 

The labelling mechanism differs according to the nature of the content: 

  • Visual and audio-visual content: The label must be prominently visible in the visual display and sufficiently noticeable and perceivable to identify the content as synthetically generated. 
  • Audio content: A prominent audio disclosure must be prefixed to the content so that listeners can identify it as synthetically generated. 

In addition to the visible or audible disclosure, the content must be embedded with permanent metadata or another appropriate technical provenance mechanism, to the extent technically feasible. This may include a unique identifier that enables identification of the content as SGI and of the computer resource of the intermediary used to create, generate, modify or alter it. 

Rule 3(3)(b) further prohibits an intermediary from enabling the modification, suppression or removal of the label, permanent metadata or unique identifier. 

The framework therefore combines user-facing disclosure with technical traceability. A visible label communicates the synthetic nature of the content to the recipient, while metadata and provenance mechanisms are intended to support identification of the resource involved in its creation or alteration. 

Importantly, the operative February 2026 Rules do not prescribe the fixed ten-percent surface-area or duration threshold that appeared in the earlier October 2025 draft. The final requirement is one of prominent visibility or audible disclosure, together with provenance measures to the extent technically feasible. 

Are the labelling requirements applicable to OTT platforms, advertisers and publishers?

The application of the Rules depends on the legal capacity in which an entity operates. The IT Rules distinguish between an intermediary and a publisher of news and current affairs content or online curated content. These are separate categories under the framework. An entity that publishes its own content does not automatically become an intermediary merely because its content is made available online. 

Consequently, a publisher of online curated content, such as an OTT service, or a news publisher that uses AI-generated material in its own editorial output should not automatically be treated as subject to the specific Rule 3(3) intermediary labelling obligation solely because it publishes synthetic content. 

Similarly, an advertiser commissioning AI-generated content is not, merely by commissioning or using that content, automatically subject to the intermediary-specific labelling obligation. 

However, this does not mean that such entities are exempt from all legal obligations relating to synthetic media. Their position may be affected by the nature of their services, whether they also operate an intermediary, the applicable provisions of the IT Rules and other laws governing the content or its publication. 

For example, a business that operates an AI-generation platform and also publishes content through that platform may need to assess its obligations in each capacity separately. An entity that merely commissions an AI-generated advertisement may have a different compliance position from the intermediary providing the generation or distribution service. The distinction is therefore not simply between social media platforms and professional publishers. It is between the different legal functions performed by the entity and the specific obligations attached to each function. 

The proposed continuous-labelling requirement: What is the current position?

The regulatory framework has also been subject to a further consultation process. In April 2026, MeitY circulated draft amendments proposing, among other changes, that the visual label required for SGI remain continuously and clearly visible throughout the duration of the content. This would be more prescriptive than the existing requirement of prominent visibility, particularly for videos and other moving visual content. 

The consultation deadline was extended to 7 May 2026. However, a draft amendment does not become operative merely because it has been published for public comments. It must be formally notified and brought into force in accordance with the prescribed process. As at 11 November 2026, the sources reviewed for this article do not establish that this proposed continuous-labelling amendment has been finally notified. Accordingly, the existing requirement under Rule 3(3)(a)(ii) should be distinguished from the proposed continuous-display standard. Intermediaries should monitor subsequent MeitY notifications and review their labelling systems if and when the proposed amendment is formally brought into force. 

Compliance considerations for businesses and platforms

The allocation of responsibility under the amended IT Rules makes it important for businesses to identify their role in the creation, modification and distribution of synthetic media. 

AI tool providers and intermediaries should consider the following measures: 

1. Assess the nature of the service: Determine whether the computer resource enables or facilitates the creation, modification, publication or dissemination of SGI. The assessment should consider the actual functionality of the service rather than relying only on its description or commercial classification. 

2. Establish a process for identifying SGI: Develop appropriate mechanisms to distinguish content that satisfies the definition of SGI from routine editing, accessibility-related uses and other expressly excluded activities. The assessment should account for the content’s apparent authenticity and its depiction of individuals or real-world events. 

3. Implement labelling and provenance mechanisms: For permitted SGI, ensure that visual content carries a prominent label and audio content includes a prefixed disclosure. Implement permanent metadata or other appropriate technical provenance mechanisms to the extent technically feasible. 

4. Prevent tampering with labels and metadata: Review product functionality to ensure that the intermediary does not enable users to remove, suppress or modify required labels, metadata or unique identifiers. 

5. Strengthen SSMI verification systems: SSMIs should establish user-declaration workflows and deploy reasonable and proportionate technical measures to verify declarations before content is displayed, uploaded or published. The verification process should take account of the nature, format and source of the content. 

6. Update user-facing policies and warnings: Review terms of service, community guidelines and user-facing notices to reflect the restrictions on unlawful SGI. Intermediaries facilitating synthetic-content creation should also provide the additional warnings required under Rule 3(1)(ca). 

7. Review grievance and escalation procedures: Ensure that systems are capable of responding to complaints involving synthetic content within the applicable timelines. The relevant teams should be able to identify complaints involving impersonation, morphed imagery, non-consensual intimate content and other categories requiring expedited action. 

8. Monitor regulatory developments: Maintain a process for tracking MeitY notifications, corrigenda, advisories and formally notified amendments. In particular, businesses should distinguish between the February 2026 requirements currently in force and subsequent proposals that remain subject to consultation. 

Conclusion

The 2026 amendments to the IT Rules establish a framework in which responsibility for synthetic media compliance depends on the role performed by the intermediary and the nature of the service it provides. The specific labelling and provenance obligations under Rule 3(3) apply to intermediaries facilitating the creation, modification or dissemination of SGI, while SSMIs are subject to additional user-declaration and technical-verification requirements before publication. 

The Rules also distinguish between the obligations of intermediaries and the responsibilities of users who create or disseminate unlawful content. Publishers and businesses commissioning synthetic media must separately assess their legal position, rather than assuming that every online entity is subject to the same intermediary-specific requirements. 

For technology companies and digital platforms, compliance will require more than adding a disclaimer to AI-generated content. It will involve identifying the relevant legal role, distinguishing SGI from excluded uses, implementing appropriate labelling and provenance systems, strengthening user-facing safeguards and maintaining effective grievance mechanisms. 

As the regulatory framework continues to evolve, businesses should ensure that their compliance programmes are based on the notified Rules and formally commenced amendments, rather than treating consultation-stage proposals as binding legal requirements. 

Last Updated on 17 November, 2026

Get King Stubb & Kasiva’s legal updates in your Google feedAdd King Stubb & Kasiva as a preferred source on Google