India's New AI Rules Are Live: Mandatory Labels, 3-Hour Takedowns

India just switched on its toughest-ever rules for AI content. From today, October 2, 2026, the amendments to the IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2026 — notified by the Ministry of Electronics and IT (MeitY) — are in force, and they fundamentally change how platforms like YouTube, Instagram, Facebook, and X handle AI-generated content.

What the new rules require

• Any AI-generated or AI-altered content that "appears to be real, authentic or true" — realistic deepfakes of people or events that are indistinguishable from reality — must carry a clear, prominent label identifying it as AI-generated or synthetic.
• Where technically feasible, platforms must embed persistent metadata or provenance identifiers into that content, so its origin stays traceable even when it is re-shared across apps.
• Once applied, those labels and metadata must not be removed, altered, or suppressed.
• Large platforms must ask uploaders to declare whether content is synthetically generated — and deploy reasonable technical measures to verify that declaration, not just take their word for it.

The 3-hour clock

• The deadline for removing unlawful AI content after a valid government or court order has been slashed from 36 hours to 3 hours.
• Complaints involving sensitive categories — non-consensual intimate imagery and impersonation — must now be actioned within 2 hours, down from 24.
• Platforms must also warn users at least once every three months about the legal penalties for misusing AI.
• Intermediaries that fail these due-diligence obligations risk losing their safe-harbor protection under Section 79 of the IT Act — the legal shield that keeps them from being liable for user content.

Why it matters

Research has found the average person identifies deepfakes correctly only about 55% of the time — barely better than a coin flip. As AI-generated audio, video, and images keep getting more realistic, mandatory labeling moves the trust question from "does this look real?" to "can this be verified?"

India is now the first major economy to pair mandatory AI labeling with takedown deadlines this tight. The EU's AI Act has transparency obligations of its own, but nothing resembling a 3-hour removal window.

AiPost's take

Expect editing apps and AI platforms serving India to rush out declaration checkboxes, watermarking, and faster moderation pipelines. The aggressive timelines will almost certainly mean more automated removals — and more false positives. Speed has a cost. But the direction is clear: the era of unlabeled synthetic media is ending, and other countries will be watching India's experiment closely.