India orders strict labels on AI-generated content from February 20, 2026

India has notified new IT Rules amendments that will force social and digital platforms to verify and clearly label AI-generated or synthetic media starting 20 February 2026.

The changes, issued by the Ministry of Electronics and Information Technology (MeitY), shift big platforms from a passive “host and remove on complaint” model to active gatekeeping of synthetic content. Significant social media intermediaries must now ask uploaders whether their content is AI-generated and use technical systems to check these claims before publication. If material is found to be synthetic, it must carry a prominent AI label and non-removable metadata or provenance markers so it can be traced back to the source service.

The rules target “synthetically generated information” such as deepfake videos, voice-cloned audio and other AI-manipulated or AI-created posts that could mislead users. Platforms are barred from letting users strip or hide these labels and can face loss of safe-harbour protections if they knowingly allow unlabeled synthetic content to spread. Separately, the government has also set a three-hour deadline for taking down flagged AI-generated or deepfake content that is objectionable or unlawful, further tightening enforcement.

Industry reaction is mixed: trust-and-safety teams warn that real-time detection at scale is technically hard and prone to false positives, while some legal and policy experts see the move as a more mature, risk-based approach compared with earlier, rigid labelling proposals like the 10% on-screen watermark rule. Commentators say the impact will be felt in product design for short-video apps, voice-cloning tools and generative AI features, which must now integrate provenance and labelling by default.

CT Bureau