Editorial illustration for India's new IT rules give Instagram and X an unworkable deepfake deadline
India Mandates Deepfake Labels on Social Media Platforms
India's new IT rules give Instagram and X an unworkable deepfake deadline
India has ordered its internet to perform a magic trick. The government's new IT rules demand that platforms like Instagram and X block deepfakes and label all AI-generated content with tamper-proof metadata. They have a deadline. What they don't have is the technology to do it.
The order assumes a level of control that Silicon Valley's own architects cannot achieve. Meta, Google, and Microsoft back the leading standard, called C2PA. It is supposed to invisibly tag synthetic media with provenance data.
The system is failing. Labels are obscure. Metadata is often absent.
A huge portion of synthetic content, from open-source models or apps that ignore the standard, carries no tags at all. Now India, with more than half a billion social media users, is telling companies to solve a problem the entire industry is stuck on.
Under India's amended Information Technology Rules, digital platforms will be required to deploy "reasonable and appropriate technical measures" to prevent their users from making or sharing illegal synthetically-generated audio and visual content, aka, deepfakes. Any such generative AI content that isn't blocked must be embedded with "permanent metadata or other appropriate technical provenance mechanisms." Specific obligations are also called out for social media platforms, such as requiring users to disclose AI-generated or edited materials, deploying tools that verify those disclosures, and prominently labeling AI content in a way that allows people to immediately identify that it's synthetic, such as adding verbal disclosures to AI audio. That's easier said than done, given how woefully underdeveloped AI detection and labelling systems currently are.
C2PA (also known as content credentials) is one of the best systems we currently have for both, and works by attaching detailed metadata to images, videos, and audio at the point of creation or editing, to invisibly describe how it was made or altered. But here's the thing: Meta, Google, Microsoft, and many other tech giants are already using C2PA, and it clearly isn't working. Some platforms like Facebook, Instagram, YouTube, and LinkedIn add labels to content flagged by the C2PA system, but those labels are difficult to spot, and some synthetic content that should carry that metadata is slipping through the cracks.
Social media platforms can't label anything that doesn't include provenance metadata to begin with, such as materials produced by open-source AI models or so-called "nudify apps" that refuse to embrace the voluntary C2PA standard. India has over 500 million social media users, according to DataReportal research shared by Reuters.
This is policymaking by aspiration. The rules command a technical fix for a problem defined by its lack of one. You cannot reliably label what you cannot definitively detect.
You cannot block what you cannot trace. India's users are already swimming in synthetic media. The tools to manage it are voluntary, inconsistent, and easily circumvented.
Expecting compliance by deadline is fantasy. The goal of curbing deepfake harm is correct. The method is pure fiction.
Real progress would require funded research, mandated standards for all AI tools, not just platforms, and a brutal acknowledgment that a label no one can see or enforce is just more data. The promise is broken before the clock even starts.
Related Reading
Policy & Regulation Grammarly faces class-action suit over AI ‘Expert Review’ feature
Policy & Regulation Claude Mythos highlights EU AI safety gaps, says researcher Caroli
Policy & Regulation After ditching AI fitness apps and a Fitbit, I return to Peloton classes
From Archives OpenAI tried to retire 4o August 2025, replace with GPT‑5 after user episodes
From Archives OpenAI Drops ‘io’ Brand After USD 6.5 B Acquisition of Jony Ive’s Hardware Unit
Common Questions Answered
What specific challenges do AI practitioners identify with MeitY's draft amendment on synthetic content?
AI practitioners argue that the draft amendment's core enforcement tools like visible labeling and platform-side verification are technically unreliable. Routine actions such as editing, compression, re-encoding, screenshots, and reposting can easily strip or degrade provenance signals, making content verification at scale practically impossible.
How do the proposed IT Rules define synthetically generated information?
The amendments define synthetically generated information as content that is artificially or algorithmically created, modified, or altered using a computer resource in a manner that appears reasonably authentic or true. This expansive definition encompasses deepfake audio, algorithmically altered photographs, fabricated metadata, and synthetically generated text.
What are the key mandatory labeling requirements in the new IT Rules draft amendment?
The amendment mandates comprehensive labeling of synthetic content, requiring intermediaries to display permanent, unique metadata. For visual content, labels must occupy a minimum of ten percent of the screen area and remain permanently visible, while audio content must have labels covering a specific portion of its duration.
Why do AI practitioners argue that the draft amendment focuses on the wrong approach to regulating synthetic content?
The practitioners argue that the draft amendment regulates how content is generated instead of addressing whether it causes harm. They warn that the technical design choices could over-regulate benign content while still failing to prevent the most prevalent forms of online deception.
Further Reading
- India orders social media platforms to take down deepfakes faster — TechCrunch
- Government's new IT rules make AI content labelling mandatory — Times of India
- India sets 3 hr deadline for social media platforms to take down AI ... — Economic Times
- Centre sets 3-hr deadline to remove flagged deepfakes — The Hans India
