Proposed amendments to the IT Intermediary Rules: Battling synthetically generated information

Introduction

On 22 October 2025, the Ministry of Electronics and Information Technology (MeitY) issued a draft notification proposing amendments (Draft Amendments) to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 (IT Rules). The Draft Amendments seek to reinforce the regulatory perimeter of the IT Rules by introducing the concept of “synthetically generated information”—a phrase intended to capture deepfakes, AI-generated, and algorithmically modified content.

The stated objective of the proposed amendment is to strengthen user safety, traceability, and accountability obligations for intermediaries—entities that receive, store, transmit, or provide services with respect to electronic records on behalf of others. The framework also covers social media intermediaries (platforms enabling online interaction and content dissemination) and significant social media intermediaries (SSMIs) (defined as platforms with more than five million registered users in India), as well as platforms that enable the creation or modification of synthetic content.

The Draft Amendments emerge against the backdrop of soaring social outrage and judicial combat with the challenges posed by deepfake and AI-generated content. Indian courts have, over the last 2 years, increasingly grappled with cases where deepfakes and other synthetically generated content have been used to impersonate individuals and facilitate fraud. These cases deal with a variety of harms, ranging from the violation of the personality rights of celebrities and business leaders (causing financial and reputational harm) to social engineering frauds on social media, and even manipulation of securities markets. Therefore, while Indian courts have, for some time now, been developing jurisprudence in the context of synthetically generated content, the draft amendment aims to regulate synthetic content in a more pro-active and structured manner.

Key Amendments Proposed — Context and Implications

1. Defining Synthetically Generated Information

  The Draft Amendments propose a definition for synthetically generated information (SGI) under a new Rule 2(1)(wa), to cover any information that is “artificially or algorithmically created, generated, modified or altered using a computer resource, in a manner that appears authentic or true.” The intent of introducing this definition is clear - to bring deepfakes and other synthetically generated content within the regulatory perimeter through a clear framework.

  However, the breadth of the proposed definition may raise as many questions as answers. A literal reading of the proposed definition of SGI would extend to any and all content that is artificially or algorithmically generated or modified – including images, video, and text. Given the nature of the safeguards sought to be introduced in respect of SGI, this will undoubtedly create some grey areas.

   The proposed definition’s use of the word “modified or altered” may also create some ambiguities, it is entirely possible to read the proposed amendment such that it includes relatively innocuous automated edits such as filters or background enhancements, AI augmented improvement of image quality, etc. Intermediaries will therefore need clarity on how to operationalise these obligations across diverse content flows.

2. Due Diligence for Platforms Enabling Creation / Modification of SGI

   The Draft Amendments intend to introduce a new Rule 3(3), requiring intermediaries that “offer a computer resource”, which could enable, permit, or facilitate the creation, generation, modification or alteration of information as SGI, to ensure that such content is labelled or embedded with a permanent unique metadata or identifier. This label must be visibly displayed (covering at least 10% of the visual surface of the SGI) or audibly marked (first 10% of audio). The intermediary must not allow removal, suppression or alteration of such labels.

  While embedding identifiers at the point of generation enhances traceability and transparency, from an operational perspective, intermediaries may face significant challenges in implementing this requirement. The 10% metric is prescriptive for visual / audio media, but its application to text or interactive AI outputs is more challenging. Platforms that integrate AI models across multiple services may need to tag each generated output, including messages and scripts. Furthermore, cross-platform content sharing, re-uploads, or subsequent edits introduce additional complexity in maintaining the integrity of the identifiers. As AI augmented devices (such as glasses) become increasingly common, a conflict may also arise between the need for a “prominent” SGI identifier, and the 10% metric - where size of screen or window-in-window displays, may mean that an SGI tag that covers 10% of the screen space is still not considered “prominent”. Ensuring compliance across diverse content formats and high-volume outputs will therefore require substantial technical, user experience, and compliance resources.

   Another question that may arise from this proposed amendment is the singling out of intermediary platforms. An online marketplace may ask why it is required to label content with SGI tags, but (for example) a car manufacturer’s website using chatbots, AI generated or augmented images and videos, etc, has to make no such declarations.

3. Enhanced Obligations for SSMIs

   Proposed new Rule 4(1A) in the Draft Amendments mandates that SSMIs are required to obtain user declarations indicating whether displayed, uploaded, or published information is synthetically generated, deploy reasonable and proportionate technical measures to verify such declarations, and ensure that SGI is clearly labelled or accompanied by a notice. Failure to carry out the above may be treated as a breach of due diligence obligations. One positive take away from this provision is that the deployment of automated tools to detect and tag SGI is restricted to such measures as are “reasonable and appropriate” – striking a balance between requiring intermediaries to deploy such tools and recognizing that the available technology will have its own limitations.

   However, from a compliance perspective, open-ended language such as “reasonable and proportionate” could create uncertainty, particularly in an industry without clear market standards, and where technology continues to rapidly evolve. Platforms would have to consider what automated tools or human review would suffice, how they document verification efforts and how they manage the volume and velocity of uploads.

4. Safe-Harbour Clarification

   The Draft Amendments clarify that intermediaries which remove or disable access to harmful synthetic content in good faith retain safe-harbour protection under Section 79(2) of the Information Technology Act 2000. However, intermediaries may face risk if synthetic content is published or circulated without labelling or verification, even where intent is not malicious. This has raised concerns that platforms may adopt an overly cautious approach, resulting in excessive removal of content or self-censorship to mitigate perceived legal risks. Platforms will need clear procedural workflows and documentation to demonstrate their good-faith efforts and to benefit from safe-harbour protection.

5.  Clarificatory Inclusion Across Obligations

   Through Rule 2(1A), the Draft Amendments clarify that any reference to “information” under rules such as 3(1)(b), 3(1)(d), 4(2) and 4(4) includes SGI. This ensures synthetic content is subject to the same removal, grievance redressal and reporting obligations as other information. However, operational issues could emerge when synthetic material is blended with authentic content. For example, if a video clip is partly real and partly AI-generated. Platforms will need to determine thresholds for when content is considered “synthetically generated” and actionable. Without clear criteria, there is a risk of over-moderation or inconsistent enforcement.

Conclusion

The Draft Amendments signal a decisive move towards pre-emptive regulation of synthetic and AI-generated content. By embedding labelling, traceability and verification duties within the IT Rules, MeitY places intermediaries squarely at the centre of compliance and enforcement. While the policy intent is clear, questions remain on definitional scope, technical feasibility and implementation standards. The consultation period therefore offers a vital opportunity for industry participants to engage with MeitY in order to ensure that the final framework achieves its true objectives without unduly constraining innovation or operational flexibility.

- Supratim Chakraborty (Partner); Madhav Khosla (Counsel) and Himeli Chatterjee (Associate)

For any queries please contact: editors@khaitanco.com