The Government of India has introduced sweeping amendments to the Information Technology Rules 2021, mandating clear labelling of AI-generated content and drastically shortening takedown timelines for unlawful material. The revised framework, notified by the Ministry of Electronics and Information Technology, will come into force on February 20, 2026. The changes are being viewed as one of the most significant regulatory interventions in India’s digital ecosystem since the original intermediary guidelines were framed.
The amendments respond to the rapid evolution of generative artificial intelligence tools that now enable users to create hyper-realistic images, videos, audio clips and even forged documents with minimal technical expertise. While such technologies offer innovation in education, entertainment, design and communication, they have also enabled deepfakes, impersonation, identity theft and targeted misinformation campaigns. Policymakers argue that regulatory guardrails are necessary to preserve trust in online spaces and protect democratic institutions.
Mandatory Labelling and Platform Accountability
Under the amended rules, social media platforms must prominently label all content that is synthetically generated or created using artificial intelligence tools. The labelling requirement applies especially to images and videos that may mislead viewers into believing they depict real events or real individuals. The intent is to ensure that users are clearly informed whenever content has been artificially generated or substantially manipulated.
Platforms with more than five million registered users will carry additional compliance responsibilities. They must obtain a formal declaration from users uploading AI-generated content confirming that the material is synthetic. In addition, such platforms are required to deploy technical verification mechanisms before publication. These measures formalize detection practices that several large technology companies already claim to use, but which are now made mandatory under law.
The government has clarified that the earlier draft definition of “synthetically generated information” was broader. The final notified version narrows its scope to avoid excessive regulation of harmless digital edits. Routine smartphone photo enhancements, automatic retouching features, colour correction tools and film industry special effects have been excluded from mandatory labelling requirements. This exemption aims to balance innovation and practicality while targeting deceptive or harmful synthetic media.
Certain categories of AI-generated content are strictly prohibited under the amended rules. These include child sexual exploitation material, forged government or legal documents, instructions related to developing explosives, and deepfakes that falsely impersonate real individuals. The inclusion of these prohibitions strengthens existing criminal safeguards and provides explicit clarity that synthetic formats do not dilute liability. Platforms hosting such content will be expected to remove it immediately and cooperate with law enforcement agencies.
Large intermediaries have also been directed to implement “reasonable and appropriate technical measures” to detect and prevent unlawful synthetic information. This includes provenance tracking systems and identifier mechanisms that allow traceability of digital media. Many global platforms are part of collaborative initiatives such as the Coalition for Content Provenance and Authenticity, which works on embedding invisible digital markers within AI-generated files. However, the Indian government has clarified that it does not mandate adoption of any single technological standard. Instead, it emphasizes the broader principle of reliable detection and traceability while allowing flexibility in implementation.
Another major shift introduced by the amendments is the strengthening of user awareness obligations. Platforms must now notify users of their terms and conditions every three months instead of once a year. These notifications must clearly explain compliance requirements, reporting obligations and consequences of violations. Users posting harmful synthetic content may face immediate removal of posts, suspension or termination of accounts, and disclosure of identity details to law enforcement authorities when required under law.
Shortened Takedown Timelines and Enforcement Mechanisms
Perhaps the most consequential change in the amended framework is the drastic reduction in takedown timelines for unlawful content. Previously, intermediaries were required to act within 24 to 36 hours upon receiving government or court orders. Under the new rules, compliance must occur within two to three hours. This accelerated window reflects the government’s position that harmful content spreads rapidly and can cause irreparable damage within a short span of time.
Timelines for user-generated complaints have also been revised. For issues such as defamation or misinformation, platforms must now respond within one week instead of the earlier two-week window. For sensitive categories of complaints under Rule 3(2)(b), the response deadline has been reduced from seventy-two hours to thirty-six hours. The official rationale behind these tighter timelines is that prolonged availability of unlawful content can amplify reputational harm, incite unrest, or compromise national security before corrective action is taken.
The government has argued that major technology platforms already possess sophisticated artificial intelligence detection tools capable of identifying manipulated media. By formalizing these obligations, the amended rules transform voluntary industry practices into enforceable legal duties. At the same time, critics have expressed concerns about feasibility, especially for platforms that handle enormous volumes of content daily. Questions remain about whether consistent compliance within two to three hours can be uniformly achieved without automated systems potentially leading to over-removal of legitimate speech.
Another area of debate relates to privacy and traceability. Provenance tracking mechanisms, while helpful in verifying authenticity, require careful safeguards to ensure they do not undermine user privacy or enable excessive surveillance. The government has attempted to address such concerns by refraining from mandating a single technological solution, thereby allowing platforms to innovate while maintaining accountability.
The amendments also signal India’s broader strategic objective of shaping global discourse on artificial intelligence governance. By combining mandatory labelling, rapid takedown obligations and stricter user accountability measures, India positions itself among jurisdictions adopting proactive regulatory frameworks to combat deepfakes and misinformation. The policy emphasis reflects recognition that artificial intelligence is no longer a niche technological tool but a foundational element of digital society.
As the rules take effect on February 20, 2026, their implementation will be closely watched by technology companies, legal experts and civil society organizations. The balance between digital innovation, freedom of expression and public safety will likely remain at the centre of ongoing policy discussions as artificial intelligence capabilities continue to evolve.
