India tightens rules on AI content, orders 3-hour takedown of flagged posts

New Delhi: In response to the growing number of complaints over AI-generated material, the Union government on Tuesday introduced stricter compliance requirements for digital platforms in dealing with synthetic content, including deepfakes. Platforms such as X and Instagram will now be required to take down such material within three hours of it being flagged by a competent authority or court.

The government has amended the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 to formally recognise and define AI-generated and synthetic content. These revised rules will take effect from February 20, 2026.

The amendments define "audio, visual or audio-visual information" and "synthetically-generated information", covering AI-created or altered content that appears real or authentic. Routine editing, accessibility improvements, and good-faith educational or design work are excluded from this definition.

One of the major changes is the classification of synthetic content as ‘information’, bringing it under the scope of unlawful activity provisions in the IT Rules. According to a gazette notification issued by the Ministry of Electronics and Information Technology (MeitY), platforms must now comply with government or court directives within three hours, a reduction from the earlier 36-hour window.

The timelines for resolving user grievances have also been shortened, PTI reported.

The amended rules mandate clear labelling of AI-generated content. Platforms that enable the creation or sharing of synthetic material must ensure it is prominently marked and embedded with permanent metadata or identifiers, wherever technically possible.

The government has also directed platforms to use automated tools to prevent the circulation of illegal, deceptive, sexually exploitative, non-consensual, or harmful AI content, including material linked to forged documents, child sexual abuse, explosives, or impersonation.

Further, intermediaries are barred from permitting the removal or suppression of AI labels or embedded metadata once applied.