Large, fixed watermarks proposed for AI content set to be dropped from IT rules

The government is preparing to notify the final amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules

The ministry of electronics and information technology (MeitY) is set to drop the requirement that would have mandated large, fixed-size watermarks on AI-generated content, two officials familiar with the matter said, as the government prepares to notify the final amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, this week.

Officials said the government is moving away from prescribing the 10% threshold.

The draft amendments released in October 2025 proposed that intermediaries offering tools to create or modify synthetically generated information must embed prominent labels or identifiers. For visual content, these labels were proposed to cover at least 10% of the surface display area, while for audio content, an audible marker was proposed during the first 10% of its duration.

For platforms such as Instagram or YouTube, this would have meant large, always visible watermarks or audio tags on every AI-generated post or video, covering part of the content and affecting the viewing experience.

Only the requirement that labels occupy at least 10% of the screen space is being dropped. Content will still need to carry a clear indication that it is AI-generated.

Officials said the ministry is moving away from prescribing the 10% threshold in the final version. “If I see a video, I should know that something is AI-generated,” one of them said.

The second official said the industry had opposed that amendment. “We have considered their request.”

The change follows pushback from industry groups that argued the rule was too rigid and technically difficult to implement across formats and devices. The Internet and Mobile Association of India (IAMAI), whose members include Google, Meta, Snapchat, WhatsApp, Zomato, Groww, Airbnb, Amazon, Apple, Jio, Airtel, Netflix, Ola, and Reddit, consulted with companies and submitted a representation to the ministry outlining their concerns.

In its submission, IAMAI said the labelling requirement was overly prescriptive and could disrupt user experience, especially for audio and video content. It noted that visible watermarks can be easily cropped or altered, metadata can be stripped, and there are no common industry standards to ensure consistent implementation. The body also flagged privacy, interoperability, and compliance burdens, particularly for smaller firms.

Industry representatives said their biggest concern remains a separate provision, Rule 4(1A), which would require significant social media intermediaries like Instagram, YouTube, Facebook, etc to obtain user declarations on whether content is synthetically generated, deploy “reasonable and proportionate” technical measures to verify those declarations, and ensure such content is clearly labelled before publication. Companies have argued that verifying declarations at scale may be difficult in practice.

The ministry, after publishing draft rules in October 2025, invited public feedback until November 6, which was extended to November 13 after requests from the industry.