Mandatory Labelling of AI-Generated Content

The Union Government has notified amendments under the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, mandating prominent labelling of photorealistic AI-generated content and reducing takedown timelines for unlawful material.

• These amendments are part of the Information Technology Rules, 2021, framed under the IT Act, 2000.

Key Amendments

• Mandatory Labelling of AI Content: Photorealistic AI-generated content must be prominently labelled to inform users that it is synthetically generated.
• Metadata Embedding: Intermediaries must embed permanent provenance markers and unique identifiers to trace content origin.
  • Platforms are prohibited from allowing removal or suppression of labels or embedded metadata
• Prohibited AI Content Categories:
  • Sexual Exploitation Content: Synthetic child sexual abuse material and non-consensual intimate imagery are strictly banned.
  • Impersonation and Fraud: AI content involving identity theft, false electronic records, or deception is prohibited.
  • Security-Sensitive Content: Synthetic media related to arms, explosives, or unlawful activities is barred.
  • Defamatory Deepfakes: Content falsely depicting individuals or events in a misleading manner is restricted.
• Reduced Timeline for Takedown:
  • Court-Declared Illegal Content: Content declared illegal by courts or government authorities must be removed within three hours
  • Sensitive Content: Deepfakes and non-consensual nudity must be taken down within two hours.
  • Grievance Redressal Reform: Acknowledgement timelines are reduced from 15 days to 7 days
  • This marks a sharp reduction from the earlier 24–36 hour window
• User Disclosure Requirement: Platforms must seek disclosures from users when content is AI-generated. In the absence of disclosure, platforms must label the content or remove it in cases of non-consensual deepfakes.
  • Platforms must deploy technical systems to verify authenticity of such declarations
• Definition: Synthetic content includes audio, visual, or audio-visual information that is artificially or algorithmically created, generated, modified, or altered in a manner that appears real or indistinguishable from natural persons or real-world events.
• Labelling Requirements:
  • Prominent Disclosure Mandate: AI-generated imagery must carry a clear and prominent label.
  • Flexibility to Platforms: Unlike the draft version, which required 10 per cent of the image to display the label, the final rules provide greater flexibility in implementation.

• Safe Harbour and Platform Liability
  • Loss of Protection for Non-Compliance: If a platform knowingly permits, promotes, or fails to act against synthetic content in violation of the rules, it may lose safe harbour protection.
  • Due Diligence Obligation: Non-compliance may be treated as a failure to exercise due diligence under the IT Rules
• Administrative Changes
  • State-Level Officers: An earlier October 2025 amendment limited each State to designating one officer for takedown orders.
  • Administrative Revision: States are now allowed to designate multiple officers, addressing administrative needs in larger States.
• Implementation:
  • The amended rules will come into force from February 20, 2026.

Significance of the Amendment

• Combatting Deepfakes and Misinformation: The rules aim to address rising concerns over non-consensual deepfakes, misinformation, and synthetic media manipulation
• Deepfake and AI Harm Regulation: The amendments represent India’s strongest legal framework yet against deepfakes and AI-driven misinformation.
• Platform Accountability: Strict labelling, metadata embedding, and rapid takedown timelines enhance intermediary responsibility.
• Cyber Governance Modernisation: Aligns India’s digital regulation with emerging AI risks while balancing safe harbour protections

About Deepfake Technology

• A deepfake is a form of digital content generated or altered through AI models, particularly Generative Adversarial Networks (GANs) and deep neural networks, that can realistically swap faces, mimic voices, alter lip movements, or fabricate entire human performances.

Check Out UPSC CSE Books

Visit PW Store

Constitutional and Legal Framework

• Freedom vs Restrictions: Regulation of deepfakes must balance free speech under Article 19(1)(a) with reasonable restrictions under Article 19(2).
• Legislative Provisions:
  • Section 69A of the IT Act empowers the government to block harmful online content following due process safeguards.
  • Section 79 grants intermediaries immunity, conditional upon due diligence and timely content removal under the safe harbour doctrine.
  • The Digital Personal Data Protection Act, 2023 mandates consent for personal data use, making non-consensual deepfakes punishable.
  • The Bharatiya Nyaya Sanhita, 2023 penalises impersonation, defamation, misinformation, and organised cyber offences linked to deepfakes
• Judicial Safeguards: Courts have emphasised proportionality, procedural fairness, and protection of constitutional rights in digital regulation.