Synthetic Content Rules 2026: Intermediary Obligations Explained

India's approach to regulating AI-generated content, deepfakes, and other forms of synthetic media has crystallised significantly in 2026 through amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules 2021. These amendments impose specific obligations on social media intermediaries and other digital service providers regarding the labelling, detection, and removal of AI-generated or manipulated content. The regulatory shift reflects global concern about the potential of deepfakes and generative AI to cause harm through misinformation, non-consensual intimate imagery, and electoral interference.

Definition and Scope of Synthetic Content

The 2026 amendments introduce a definition of synthetic content that encompasses images, audio, video, and text generated or substantially manipulated using AI models such that a reasonable person could not readily identify the content as artificial without disclosure. The definition is deliberately broad to capture technology developments rather than being tied to any particular technique. The rules distinguish between clearly labelled AI-generated creative content and deceptive synthetic content, with obligations applying most stringently to the latter. The rules apply to content hosted or transmitted by intermediaries operating in India, regardless of where the AI generation took place.

Labelling and Disclosure Obligations

Significant intermediaries, defined as those with over five million registered users in India, are required to implement technical measures to detect AI-generated content and display clear labels on such content before it is visible to users. Where detection is not technically feasible, intermediaries must require users who upload synthetic content to disclose its AI-generated nature at upload. Platforms must provide users with a mechanism to report unlabelled synthetic content and must process reports within a prescribed timeframe. The obligations are particularly stringent for synthetic content depicting real persons without consent.

Takedown Timelines and Penalty Framework

For deepfakes depicting real persons in sexually explicit contexts without their consent, platforms must act within 24 hours. For synthetic content constituting election misinformation during notified election periods, takedown must occur within three hours. Other categories of harmful synthetic content attract a 48-hour window. Platforms that fail to comply lose their safe harbour protections under Section 79 of the IT Act for the specific content, and repeat failures attract escalating penalties and blocking orders.

Practical Takeaways

Technology companies operating social media, video sharing, gaming, or communication platforms in India must assess whether they qualify as significant intermediaries and implement AI content detection and labelling systems meeting the 2026 requirements. Legal and policy teams should update content governance frameworks, grievance redressal systems, and user terms of service. Businesses using generative AI for marketing content should ensure any content depicting real persons is clearly labelled and that required consents have been obtained. Regular review of MEITY guidance and enforcement actions is essential in this fast-moving area.