Regulation of AI Content in India

• The Union Government notified amendments to the IT (Intermediary Guidelines & Digital Media Ethics Code) Rules, 2021, mandating labelling of AI-generated content.

Key Amendments Notified

• Mandatory AI Labelling: Photorealistic AI-generated or synthetically generated content must carry a prominent disclosure, preventing users from mistaking it as real.
• Compressed Takedown Timelines: Platforms must remove court/government-flagged unlawful content within 3 hours and non-consensual deepfakes within 2 hours, reduced from 24–36 hours.
• User Disclosure Duty: Platforms must seek self-declaration from users on whether content is AI-generated, failing which platforms must label or remove it.
• Narrowed AI Definition: Routine edits and quality-enhancing AI tools (e.g., camera touch-ups) are excluded from the definition of synthetically generated content.

Rationale Behind the Amendment

• Prevent Viral Harm: Unlawful and deepfake content often goes viral within minutes; studies show over 60% of harmful content reaches peak circulation within 6 hours.
• Protect Dignity and Privacy: India has seen a sharp rise in non-consensual intimate imagery (NCII) cases; NCRB data shows cybercrime cases rose by over 31% between 2022–2023.
• Platform Responsibility: With India having over 850 million internet users, intermediaries are now expected to exercise higher due diligence proportional to their technological capacity.
• Ethical AI Governance: Aligns with OECD AI Principles and G20 AI Safety Guidelines to promote responsible AI deployment in India.

Concerns and Challenges

• Operational Feasibility: A 2–3 hour takedown window may be difficult where illegality is ambiguous, especially when law enforcement notices lack detailed reasoning.
• Over-Censorship Risk: Fear of penalties and loss of safe harbour may push platforms towards precautionary takedowns, chilling legitimate speech and satire.
• Safe Harbour Uncertainty: Non-compliance can trigger loss of intermediary immunity, exposing platforms to criminal and civil liability for user-generated content.
• Compliance Burden: Smaller platforms and start-ups may lack real-time AI detection tools and moderation staff, creating uneven regulatory impact.

Be Prelims Ready!

Way Forward

• Clearer Illegality Tests: Issue standardised, content-specific guidelines to identify unlawful and deepfake material; E.g., predefined indicators for NCII, impersonation and election-related misinformation.
• Graded Timelines: Introduce risk-based takedown timelines instead of a uniform window; E.g., immediate removal for NCII, extended review for context-dependent speech.
• Independent Oversight: Establish an independent appellate or review mechanism to address wrongful takedowns; E.g., a digital content ombudsman with time-bound decisions.
• Tech Enablement: Support platforms through shared AI detection tools, hash databases and government-backed infrastructure; E.g., a national deepfake detection and verification facility.

“Technology must be governed, not gagged.” While the amended IT Rules respond decisively to AI-driven harms, their democratic credibility rests on clear standards, proportional enforcement, and independent oversight to protect constitutional freedoms.

Reference: The Hindu | PMFIAS: IT Amendment Rules, 2025

PMF IAS Pathfinder for Mains – Question 544

Q. The amended IT Rules, 2021, impose enhanced due diligence and compliance obligations on intermediaries. Examine whether these measures are adequate to regulate AI-generated content without infringing constitutional freedoms. (250 Words) (15 Marks)

Approach

• Introduction: Write a brief introduction about the amended IT Rules, 2021.
• Body: Write about the adequacy of the measures to regulate AI-generated content, mentioning concerns over constitutional freedoms and the way forward.

• Conclusion: Emphasis on a rights-respecting framework to balance innovation with constitutional freedoms.