Govt tightens AI deepfake rules | Cuts unlawful content takedown timeline from 36 to 3 hours

The Centre on Thursday unveiled a strengthened regulatory framework to tackle the growing threat of AI-generated deepfakes, introducing stricter obligations for social media intermediaries, mandatory labelling of AI-generated content and sharply reducing the timeline for removing unlawful material from online platforms.

The Ministry of Electronics and Information Technology (MeitY) said the revised framework, notified through amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, seeks to address harms arising from synthetically generated information (SGI), including deepfakes and other AI-generated content.

One of the key changes is the reduction in the timeline for removal of unlawful content after valid government notices or court orders—from 36 hours to 3 hours. Timelines for grievance redressal have also been shortened, with complaints involving sensitive categories such as nudity and impersonation now required to be addressed within 2 hours, down from 24 hours.

AI-generated content must be labelled

Under the amended rules, intermediaries will be required to ensure that permissible AI-generated content is clearly labelled and carries traceable metadata, enabling users to identify synthetic material and reducing the risk of deception. The government said the amendments also strengthen user accountability and platform due diligence by mandating awareness of the legal consequences of creating or sharing unlawful AI-generated content.

The framework explicitly covers harmful content such as child sexual exploitation material, non-consensual intimate imagery, impersonation and other malicious AI-generated media, requiring platforms to detect and remove such content promptly.

Platforms face stricter compliance

The amended rules require intermediaries to deploy appropriate technical measures, including automated tools wherever necessary, to prevent users from generating, publishing or disseminating AI-generated content that violates Indian law. Significant Social Media Intermediaries (SSMIs) will be expected to proactively identify unlawful content through automated detection systems, while continuing to publish compliance reports, appoint grievance officers and facilitate user appeals.

The government warned that intermediaries failing to comply with their legal obligations could lose the "safe harbour" protection available under Section 79 of the Information Technology Act, exposing them to legal action under applicable laws.

IndiaAI Mission backs indigenous deepfake detection

Highlighting efforts under the IndiaAI Mission, the ministry said the Safe and Trusted AI pillar has already approved 13 Responsible AI projects across educational institutions, including several focused on deepfake detection. Among the major initiatives are Saakshya, a multi-agent deepfake detection framework developed by IIT Jodhpur and IIT Madras; AI Vishleshak, aimed at improving audio-visual forgery detection; and IIT Kharagpur's Real-Time Voice Deepfake Detection System. The government said these projects are intended to strengthen indigenous AI governance tools and promote responsible deployment of artificial intelligence.

Multiple agencies to coordinate enforcement

The Centre said enforcement against unlawful AI-generated content would involve multiple agencies and mechanisms, including the Indian Cyber Crime Coordination Centre (I4C), the SAHYOG portal, the National Cyber Crime Reporting Portal, state police forces and the Samnvaya platform for cyber fraud investigations. Citizens can report deepfakes and related cyber offences through the National Cyber Crime Reporting Portal or by calling the cybercrime helpline 1930. The Grievance Appellate Committees (GACs) will continue to serve as appellate forums for users dissatisfied with the decisions of intermediaries regarding content moderation.

Existing laws also apply

The government emphasised that deepfakes are already punishable under multiple provisions of the Information Technology Act, 2000, and the Bharatiya Nyaya Sanhita (BNS), depending on the nature of the offence. These include provisions relating to identity theft, impersonation, forgery, misinformation, obscenity, privacy violations and organised cybercrime. The ministry said the amendments form part of a broader AI governance framework aimed at ensuring that artificial intelligence is developed and deployed in a safe, trusted and accountable manner while safeguarding users from emerging digital threats.