The Ministry of Electronics and Information Technology (MeitY) has recently proposed several changes to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 (the “IT Rules”), including the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Second Amendment Rules, 2026 released on April 21, 2026 (the “Proposed April Rules”). We have, in a previous article, discussed the changes to the IT Rules previously proposed in March 2026 (the “Proposed March Rules”), including expanded responsibilities of individual users as creators of online news content. The Proposed April Rules contain the same proposed amendments as the Proposed March Rules, save for one additional inclusion.
Under the IT Rules, “synthetically generated information” includes visual or audiovisual information that is artificially created, generated, modified or altered in a way that appears to be real, and depicts an individual in a manner that is (or is likely to be perceived as) real. Under the Proposed April Rules, all synthetically generated information is required to be labelled in a way that, “ensures continuous and clearly visible display of such label throughout the duration of the content, in a visual display.” This is a departure from the current regulations, that merely require the labelling to be prominently visible in the visual display.
The latest iteration of the IT Rules and the inclusion of “synthetically generated information” as part of its ambit has not been free of criticism, the chief amongst which is that it is overbroad and sweeps far more content into its ambit than perhaps intended. Another key criticism has been of its new labelling requirements. The IT Rules now also require all pieces of synthetically generated information to be subject to heightened due diligence obligations by intermediaries, including by embedding metadata to enable identification of the originator, and deploying technical measures to prevent any user from creating any synthetically generated information that falsifies an electronic record or document. Aside from the risk of over-correction and the automatic removal of several pieces of legitimate content, a key concern is also regarding over-labelling and the understanding of synthetically generated information.
Though exceptions to regulated synthetically generated information have been spelled out in the IT Rules, including good faith editing, colour correction, and enhancement if it does not materially alter, distort, or misrepresent the substance, context, or meaning of the underlying information. Therefore, in order to determine whether a piece of information or content should be treated as synthetically generated information or whether it falls under the exception requires a degree of determination and shouldn’t be left up to the discretion of automated content filters.
Then comes the issue of labelling. The use of AI-enabled tools to create or edit content is nothing new. Take the proliferation of short form AI-generated humanoid-fruit videos with narrative storylines. It would be absurd to require the labelling of content so obviously false. A musician using an app like Suno or technology like autotune may also find themselves in deep water. If T-Pain were to preface every one of his songs with a “prominently prefixed audio disclosure,” it is hard to see how he would have very many listeners despite his droves of talent. What about an Instagram influencer using Facetune to smoothen out a blemish or removing red-eye from a picture of a night out? Requiring them to disclose that the image is synthetically generated because of minor editing could have more significant implications for their individual brand and for the creator economy.
While the prominent labelling requirement under the IT Rules certainly wasn’t perfect, the proposed changes in the Proposed April Rules take it a step further than required. Though we may all benefit from a little more transparency in the online world, requiring every piece of synthetically generated visual information to be continuously and clearly labelled throughout the duration of its display is excessive. Between prominent and clear and continuous labelling, we are facing a new reality where a majority of the content we interact with online will likely have a ‘synthetically generated’ label, which is sure to hamper our doomscrolling experience.
English
Chinese (Simplified)
Japanese