Deepfake Watermarking Mandate — Main Idea

A deepfake watermarking mandate means requiring AI-generated or AI-manipulated audio, images, or videos to carry a visible label and/or machine-readable provenance marker showing that the content is synthetically generated. The purpose is to let users distinguish authentic media from manipulated media and to make tracing the origin of synthetic content easier.

Position in India

India has moved toward mandatory labelling and traceability of “synthetically generated information” (SGI) through the 2026 amendments to the IT Rules. The framework requires platforms to use technical measures for identifying synthetic content and to make disclosures sufficiently prominent for users. The Government has described the amendments as addressing deepfakes and other AI-generated content.

Strictly speaking, therefore, “watermarking” should not be understood as merely placing an “AI-generated” logo on a video. The regulatory approach also involves technical identification/traceability and user-facing disclosure.

Important case laws

1. Narendra Kumar Goswami v. Union of India (2025)
A PIL specifically sought directions requiring watermarking of AI-generated content, including metadata identifying its origin, tools used and creator. The case is significant because it directly raised the question of whether India should impose a statutory AI-watermarking regime.

2. Anil Kapoor v. Simply Life India & Ors. (Delhi HC, 2023)
The Delhi High Court restrained unauthorized use of Anil Kapoor's name, image, voice, likeness and personality through AI, deepfakes and face-morphing technologies. The Court recognized the connection between personality rights, privacy and dignity.

3. Mr Sudhir Chaudhary v. Meta Platforms Inc. & Ors. (Delhi HC, 2025)
The Court dealt with AI-generated deepfake videos falsely depicting the journalist and held that his name, likeness and voice had been unlawfully used to create misleading content. This demonstrates why identification and rapid removal mechanisms are legally important.

4. Raghav Chadha v. Ashok Kumar/John Doe & Ors. (Delhi HC, 2026)
The Court considered deepfakes and synthetic media involving a person's identity, including voice cloning and AI-generated videos. It emphasized the need to balance protection of dignity and personality rights with freedom of speech and expression.

One example

Suppose an AI tool creates a realistic video showing a famous actor promoting a fake investment scheme. A watermark/label such as “Synthetically Generated”, together with machine-readable provenance information, can alert viewers that the video is not authentic. If the video nevertheless impersonates the actor and causes harm, courts can additionally grant injunctions, takedown orders and protection of personality rights, as seen in Anil Kapoor and later deepfake cases.

Short recap

Deepfake watermarking mandate = transparency + traceability + accountability.

The basic legal idea is: AI-generated content should be identifiable as synthetic, while harmful impersonation can additionally attract takedown orders, injunctions and personality/privacy remedies. India's 2026 IT Rules significantly strengthen this approach.