AI Content Must Be Labelled, Deepfakes to Be Removed in 3 Hours: New IT Rules
New Delhi, 10th February 2026: Social media platforms such as X (formerly Twitter), YouTube, Snapchat and Facebook will now be required to clearly label Artificial Intelligence (AI)-generated content and remove deepfake videos and images within three hours of being reported, under new rules notified by the Ministry of Electronics and Information Technology (MeitY).
The ministry issued the order on Tuesday, amending the Information Technology (IT) Rules, 2021. The revised rules will come into force from February 20. The draft of these amendments was released by the government on October 22, 2025.
The new framework aims to curb the misuse of AI and deepfake technology by making it mandatory to label and trace synthetic content. Any AI-generated audio, video, or visual material must now clearly mention that it is not real and has been created using AI tools. The move is expected to reduce misinformation, online fraud and election manipulation.
Mandatory Labels and Metadata
Under the amended Rule 3(3), social media intermediaries that allow users to create or share “synthetically generated information” must ensure that such content carries a prominent label. Platforms will also have to embed permanent, unique metadata or identifiers in every AI-generated post.
The label must cover at least 10 percent of the visual area in images or videos, or be played during the first 10 percent of the duration in audio content. The embedded metadata cannot be altered, hidden, or deleted. Platforms will also be required to adopt technical systems to detect AI-generated material before it is uploaded.
The move follows several high-profile deepfake incidents, including a viral manipulated video of actor Rashmika Mandanna and another of cricket legend Sachin Tendulkar allegedly promoting a gaming app.
Three Key Changes in the IT Rules
First, social media companies will no longer be allowed to remove or conceal AI labels or metadata once applied. Second, platforms must use automated tools to block illegal, obscene or fraudulent AI-generated content. Third, intermediaries must issue warnings to users at least once every three months, informing them of penalties or legal action for misuse of AI or violation of the rules.
Impact on Users and Industry
The new rules will help users identify fake content more easily and reduce the spread of misinformation. However, content creators will need to take extra steps to ensure proper labeling. For platforms, the challenge will be investing in metadata systems and verification technologies, which could increase operational costs.
Despite these challenges, the government believes the rules will play a key role in preventing the misuse of AI.
What the Ministry Said
The IT Ministry said the move is aimed at building an “open, safe, trusted and accountable internet” and addressing risks such as misinformation, impersonation and election interference driven by generative AI.
What Is a Deepfake?
A deepfake is a manipulated video, image or audio clip in which a person’s face, voice or expressions are altered using AI tools. The editing is often so realistic that it becomes difficult to distinguish fake content from real material.
