Govt Proposes Stricter AI Labelling Norms

Govt Proposes Stricter AI Labelling Norms

Why in News?

Recently, the Ministry of Electronics and Information Technology (MeitY) proposed stricter amendments to the IT Rules, 2021, mandating all AI-generated content to carry clearly visible labels throughout its entire duration.

Highlights of Govt Stricter AI Labelling Proposal 

• Continuous AI Labelling Mandate: The Government of India (MeitY) in April 2026 proposed a single key norm—mandatory continuous labelling of AI-generated content. 
  • Under draft amendments the platforms must ensure a “continuous and clearly visible” label on all synthetic or AI-generated content throughout its entire duration.
  • This proposal replaces the earlier requirement of “prominent visibility” with a stricter persistent disclosure mechanism. 
  • The label can no longer appear briefly (start/end) but must remain visible at all times, marking a significant regulatory tightening in India’s AI governance framework. 
• Legal Basis: The change is introduced under amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, specifically modifying provisions under Rule 3 obligations for intermediaries. 
  • The focus is on strengthening due diligence and transparency obligations related to synthetically generated information (SGI).
• Scope of Applicability: The norm applies across all digital formats—including text, images, audio, and video content generated or significantly modified using AI tools. 
  • It also covers synthetically altered media (e.g., deepfakes, voice clones, manipulated visuals), ensuring universal and format-neutral compliance.
• Technical Requirement: The proposal mandates persistent on-screen labels or watermarks, meaning users must see the AI disclosure throughout playback or viewing, even during scrolling or interaction. 
  • This eliminates the risk of users missing disclaimers due to platform design or short attention spans. 
• Public Consultation: The government has extended stakeholder and public consultation deadline to 7 May 2026, allowing inputs from industry, civil society, and citizens before final notification.

Why Continuous AI Labelling Is Needed?

• Rising Deepfake Threat: The primary need arises from the rapid growth of deepfake and synthetic media misuse. 
  • AI tools can create highly realistic fake videos, audio, and images, often indistinguishable from real content.
  • Government data highlights increasing cases of impersonation, misinformation, and fraud, making continuous labelling essential.
• Misinformation at Scale: India has nearly 1 billion internet users, making it highly vulnerable to viral misinformation spread through AI content. 
  • Fake content can manipulate elections, incite social unrest, and distort public opinion, necessitating clear, always-visible labels to ensure informed digital consumption. 
• Inadequacy of Earlier Norms: Earlier rules required only “prominent” or partial visibility (10% display/duration) of labels, which users could easily miss. This created information gaps. 
  • Hence, the shift to continuous labelling ensures 100% visibility throughout content consumption. 
• Protection of Individuals’ Rights: Unlabelled AI content has led to fake endorsements, reputational damage, and identity misuse, especially of public figures.
  • Labelling helps prevent wrongful association and personality rights violations, strengthening digital identity protection frameworks in India.
• Platform Accountability Gap: Digital platforms previously operated with limited liability under safe harbour provisions. With AI misuse rising, the government identified the need to impose stronger due diligence and transparency obligations.

AI Governance Framework in India

• Foundational Principles (Sutras): Since 2018, India has followed the “AI for All” philosophy. 
  • The November 2025 India AI Governance Guidelines formalised seven “Sutras”, including Trust, Human-Centricity, and Accountability.
  • These principles ensure that responsible innovation is prioritised over cautionary restraint, aligning with the vision of Viksit Bharat 2047. 
  • These guide the IndiaAI Mission, which has an outlay exceeding ₹10,300 crore to build a trusted AI ecosystem.
• Institutional Oversight Bodies: A “whole-of-government” approach is led by the AI Governance Group (AIGG) for policy coordination and the Technology & Policy Expert Committee (TPEC) for advisory.
  • Additionally, the IndiaAI Safety Institute (AISI) provides technical validation and safety benchmarking. 
  • Indian Cyber Crime Coordination Centre (I4C) actively monitors AI-driven cyber threats.
  • In November 2019, the Government set up a Fact Check Unit (FCU) under the Press Information Bureau (PIB), Ministry of Information and Broadcasting. 
• Information Technology Act, 2000: The Information Technology Act, 2000 remains India’s primary digital legislation, governing e-commerce and cyber activities. 
  • It penalizes offenses like unauthorized access, data theft, and hacking under Section 43/66, and punishes identity theft/impersonation via Section 66C/D, ensuring secure digital transactions.
• IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021: The IT Rules, 2021, remain the primary legislation governing digital ecosystems, mandated under Section 87 of the IT Act, 2000. 
  • They enforce intermediary accountability (Rule 3) to tackle crimes like child pornography, illegal acts, and misinformation, requiring prompt removal.
• Data Privacy and AI Training: The Digital Personal Data Protection (DPDP) Act, 2023, serves as the backbone for AI data governance. 
  • It mandates explicit, informed consent for processing personal data used to train AI models. 
  • Significant Data Fiduciaries (SDFs) handling high-volume AI data must appoint an India-based Data Protection Officer and conduct periodic Data Protection Impact Assessments. 
• Synthetically Generated Information (SGI): The IT Amendment Rules, 2026 (effective 20 February 2026) introduced the legal category of Synthetically Generated Information (SGI). 
  • This covers AI-created text, audio, and video that appear authentic. 
  • All intermediaries must now ensure prominent and visible labelling of SGI to prevent deception. 
• Hyper-Fast Takedown Mandates: Compliance timelines have been drastically compressed. Intermediaries must remove unlawful SGI within 3 hours of receiving a government or court order. 
  • For high-sensitivity content like non-consensual deepfake nudity, the takedown window is further slashed to just 2 hours to curb viral spread.
• Metadata and Provenance Tracking: Transparency is enforced through technical “Digital DNA.” Platforms are required to embed permanent metadata and unique identifiers in AI files. 
  • These markers must be tamper-proof, allowing authorities to trace the original computer resource and device used for content creation.
• Platform Accountability and Safe Harbour: Significant Social Media Intermediaries (SSMIs)—those with over 5 million users—must now obtain user declarations for AI content. 
  • Intermediaries must block synthetic content involving Child Sexual Abuse Material (CSAM), non-consensual intimate imagery (NCII), false documents, or deceptive impersonation etc.
  • Failure to deploy automated verification tools or follow the new timelines results in the loss of Safe Harbour protection under Section 79 of the IT Act, exposing platforms to direct civil and criminal liability.
• Quarterly Compliance Education: User awareness is no longer annual. Intermediaries must notify users of platform rules and legal consequences (including penalties under the Bharatiya Nyaya Sanhita, 2023) at least once every 3 months. 
  • This ensures “Digital Nagriks” are consistently informed of their rights and duties in the AI era.