Listen to this Post
India has stepped into a new regulatory era. For the first time, artificial intelligence generated content is no longer operating in a grey zone. On February 10, the central government issued a formal notification amending the Information Technology Intermediary Guidelines and Digital Media Ethics Code Rules, 2021, bringing AI generated material under direct legal oversight. The amendments, notified as G.S.R. 120(E) and signed by MeitY Joint Secretary Ajit Kumar, will take effect from February 20. The move signals a decisive shift from voluntary labeling and platform driven moderation to structured legal accountability.
AI Content Now Defined Under Law With Clear Boundaries
The new rules introduce a formal definition of “synthetically generated information” or SGI. This includes any audio, visual, or audiovisual material that is artificially or algorithmically created, modified, or altered using computer resources. The defining feature is not just that a machine was involved, but that the content appears real or authentic and could reasonably be mistaken as genuine.
This definition covers deepfake videos, AI generated voiceovers, face swapped images, and manipulated visuals that portray real individuals or events. Even fictional scenarios involving real public figures may fall within scope if they convincingly simulate reality. The emphasis is on deception risk. If the content can blur the line between real and fabricated, it qualifies.
Routine Editing and Conceptual Work Are Exempted
The government has carved out important exceptions. Standard edits such as color correction, compression, transcription, translation, accessibility enhancements, and noise reduction are not considered SGI as long as they do not alter the meaning of the original material. Similarly, illustrative or conceptual visuals used in research papers, presentations, PDFs, or training material are excluded.
Content created purely for hypothetical, draft, template based, or conceptual purposes is also outside the definition. An office presentation with an AI generated stock illustration remains permissible. A deepfake speech of a political leader delivering words never spoken does not.
Mandatory Labels Must Be Clear and Visible
The most immediate impact is on platforms. Any intermediary that enables or facilitates the creation or spread of SGI must now label it clearly and prominently. The disclosure cannot be hidden in metadata or buried in obscure corners of a platform interface. It must be visible on the content itself in an unambiguous manner.
Beyond visual labeling, platforms are required to embed persistent metadata and unique identifiers into synthetic content, where technically feasible. These identifiers must allow traceability back to the intermediary’s system. Once embedded, platforms are prohibited from enabling removal or tampering. This provision directly addresses the long standing loophole where labels disappeared once content was downloaded and re uploaded.
Significant Social Media Platforms Face Stricter Compliance
Large platforms such as Instagram, YouTube, and Facebook, categorized as significant social media intermediaries, face additional obligations under the newly introduced Rule 4(1A). Before any upload goes live, users may be required to declare whether their content is synthetically generated.
Platforms must then deploy automated tools to verify those declarations. If AI generation is confirmed, the content must carry a visible label or notice before being made accessible to viewers. Failure to enforce this framework carries serious consequences. If a platform knowingly permits or promotes unlabeled synthetic content, it may be considered to have failed due diligence. That failure risks the loss of safe harbour protections, exposing companies to legal liability.
Watermark Size Proposal Dropped After Industry Pushback
An earlier draft released in October 2025 proposed strict watermark requirements. Visual labels were to occupy at least 10 percent of display area, while audio disclosures would have played during the first 10 percent of a clip. Industry groups such as IAMAI argued that these rigid thresholds were impractical and technologically burdensome.
The final rules drop that quantitative requirement. Labels remain mandatory, but platforms have flexibility in presentation. This signals a regulatory compromise that balances enforcement with operational feasibility.
Takedown Timelines Reduced Dramatically
The amendments also tighten response timelines. In certain cases involving lawful government orders, platforms now have just three hours to act, reduced from the earlier 36 hour window. Other compliance periods have been shortened from 15 days to seven days and from 24 hours to 12 hours.
Platforms must proactively deploy automated tools to block synthetic content that violates the law. The categories explicitly mentioned include child sexual abuse material, obscene or pornographic content, false electronic records, content related to explosives or weapons, and deepfakes that intentionally misrepresent real individuals or events.
Mandatory User Warnings Every Three Months
The obligations extend to user awareness. Platforms must warn users at least once every three months about the consequences of misusing AI generated content. These warnings may be delivered through terms of service updates, privacy policy revisions, or in app notifications. They must be available in English or any language listed in the Eighth Schedule of the Constitution.
Consequences for misuse include account suspension, termination, and potential reporting to law enforcement under the Bharatiya Nyaya Sanhita or the Protection of Children from Sexual Offences Act. References to the Indian Penal Code have now been updated to reflect the Bharatiya Nyaya Sanhita, 2023, aligning the digital framework with India’s revised criminal law architecture.
What Changes for Everyday Users
For ordinary social media users, the most visible shift will be labeling. AI generated reels, videos, voice clips, and images will carry disclosure tags. Before liking, sharing, or forwarding content, users will increasingly see notices indicating that the material was machine created or modified.
Upload flows may also include declaration steps requiring users to confirm whether AI tools were used. False declarations are no longer minor policy violations. Depending on context, they may attract legal penalties under applicable criminal statutes.
Platforms must also regularly remind users about AI content rules and associated liabilities. These reminders will become a recurring feature of the digital ecosystem rather than a one time announcement.
What Undercode Say:
India’s move is not merely about labeling deepfakes. It represents a structural shift in how digital authenticity is treated in law. By defining synthetically generated information in terms of its potential to mislead, the government has placed perception at the center of regulation. This is significant. The harm of AI content is rarely about the technology itself. It is about its capacity to convincingly simulate reality.
The requirement for persistent metadata is arguably the most powerful element of the amendments. Labels can be cropped. Watermarks can be blurred. But embedded identifiers that survive platform transfers and file compression introduce a forensic layer to digital content. This could reshape investigations into misinformation campaigns, financial fraud, and reputational attacks.
The shortened takedown timelines suggest that the state anticipates high velocity harm. In the age of viral content, 36 hours can be an eternity. Reducing that window to three hours reflects an understanding that synthetic misinformation can influence elections, markets, or public order within minutes. The law is adapting to the speed of algorithms.
At the same time, enforcement complexity cannot be ignored. Automated verification of user declarations requires sophisticated AI detection systems. Current detection tools are imperfect. False positives could burden creators who use legitimate editing tools. False negatives could allow high quality deepfakes to slip through. Platforms now carry the dual burden of compliance and technical accuracy.
There is also a strategic signaling component. India positions itself as a major digital economy with expanding AI ambitions. Regulating AI generated content while encouraging AI innovation creates a delicate balance. Too strict a regime could chill experimentation. Too lenient an approach could invite digital chaos. The removal of rigid watermark percentages shows regulatory pragmatism.
The explicit integration of the Bharatiya Nyaya Sanhita signals another dimension. AI misuse is no longer treated as a niche cyber issue. It is being folded into mainstream criminal jurisprudence. This elevates digital manipulation from platform policy violation to potential criminal offense, especially in cases involving exploitation, impersonation, or national security risks.
From a geopolitical perspective, India’s approach aligns with a broader global trend toward AI accountability. Jurisdictions worldwide are grappling with deepfakes and synthetic misinformation. By formalizing definitions and imposing traceability obligations, India enters the emerging international conversation about digital provenance and authenticity standards.
Yet, regulatory success will depend on implementation clarity. Terms like “appears real” or “could be mistaken for genuine” invite interpretative challenges. Courts may eventually need to define thresholds of deception. The evolution of jurisprudence around synthetic content will shape how flexible or rigid the framework becomes.
Ultimately, the amendments shift responsibility from users alone to platforms as gatekeepers. Safe harbour protections are now linked directly to AI labeling diligence. This recalibrates platform incentives. Compliance is no longer optional reputation management. It is structural risk mitigation.
Fact Checker Results
✅ The amendments formally define synthetically generated information and require clear labeling by intermediaries.
✅ Takedown timelines for certain government orders have been reduced to as little as three hours.
❌ The final rules do not mandate a fixed 10 percent watermark size, that proposal was removed from the draft.
Prediction
AI labeling will become a normalized digital feature across major Indian platforms within months. 📊
Legal disputes over what qualifies as deceptive synthetic content are likely to increase as enforcement begins. ⚖️
India’s framework may influence other emerging markets seeking balanced AI regulation models. 🌍
▶️ Related Video (82% Match):
🕵️📝✔️Let’s dive deep and fact‑check.
References:
Reported By: timesofindia.indiatimes.com
Extra Source Hub (Possible Sources for article):
https://www.digitaltrends.com
Wikipedia
OpenAi & Undercode AI
Image Source:
Unsplash
Undercode AI DI v2
Bing
🔐JOIN OUR CYBER WORLD [ CVE News • HackMonitor • UndercodeNews ]
📢 Follow UndercodeNews & Stay Tuned:
𝕏 formerly Twitter 🐦 | @ Threads | 🔗 Linkedin | 🦋BlueSky | 🐘Mastodon
