India tightens AI deepfake rules, orders 3-hour takedown window
India tightens AI deepfake rules, orders 3-hour takedown window
New IT rules require clear labeling of AI-generated content and a three-hour removal window for flagged material, redefining how platforms handle synthetic media.
A sweeping change to India's digital rules is taking effect soon as the government lays down new obligations for social platforms and other intermediaries. The amendments to the Information Technology Rules 2021 define AI-generated content and set a tight three-hour window to remove or disable access to material flagged as objectionable by a competent authority or court. The changes are scheduled to come into force on February 20 and are issued under Section 87 of the IT Act, 2000.
Under the new regime, platforms must clearly label AI-generated content and ensure verifiable metadata is embedded to help trace origins and edits. The scope covers intermediaries including social media platforms, messaging apps, video-sharing platforms, and other AI-driven content services that host or enable access to user-generated content. In addition to labeling, platforms will need to verify user declarations on AI content and implement traceable metadata to support enforcement.
The rules also define what counts as AI-generated or synthetic content, including audio, visual, or audiovisual content created or altered using computer resources in a way that can appear authentic. Routine editing such as colour correction, noise reduction, compression and translation is exempt, provided the original meaning remains intact. Researchers’ papers, training materials, PDFs and illustrative drafts are likewise not covered by these provisions.
A three-hour takedown deadline introduces a new pressure point for platforms to act swiftly when content is flagged as objectionable. Penalties for non-compliance or delays remain a key consideration as authorities prepare to implement the framework. Industry observers expect the rules to influence how synthetic media is disseminated in India, potentially curbing harmful deepfakes while prompting discussions about innovation and freedom of expression as the February 20 start date nears.
As platforms adapt their moderation workflows and user communication strategies, the move signals a broader push toward digital safety and accountability in the rapidly evolving AI landscape.
