Particle.news
Download on the App Store
30 ARTICLES
·
2mo ago

India Enforces Rapid Takedowns and AI Labelling as It Presses Meta for Algorithm Fixes

Tighter IT rules require visible labels and traceable metadata for synthetic content, three‑hour removals and a Meta compliance roadmap as regulators consider removing legal safe‑harbour.

The government seeks to speed up the detection and removal of deepfakes, morphed images and misleading posts.
Govt Tells Meta To Overhaul Algorithms, Submit Roadmap Against Deepfakes & Propaganda
The government revised the IT rules in February.
Image

Overview

  • MeitY announced on Thursday that platforms must remove unlawful content within three hours of a government or court notice and must act within two hours for sensitive categories such as nudity and impersonation.
  • The amended IT rules require clear visible labels and traceable metadata on AI‑generated or synthetically altered text, audio and video so users can identify manipulated material.
  • Significant Social Media Intermediaries with over 50 lakh users must deploy automated detection tools, publish compliance reports, appoint grievance officers and enable verification and appeals.
  • The government held multi‑day technical and policy meetings with Meta, pressed the company for specific algorithmic changes and has asked for a detailed compliance roadmap while reviewing whether Meta still qualifies for Section 79 safe‑harbour.
  • New investments and projects under the IndiaAI Mission and at IITs will build domestic deepfake and audio‑visual forgery detection capacity to support faster detection and enforcement.