Examine the adequacy of India's legal framework against non-consensual and morphed imagery of women.

Non-consensual intimate imagery (NCII), including morphed photographs, is a form of gender-based online abuse that strikes at a woman's dignity under Article 21. India's response rests on conditional safe harbour for intermediaries under Section 79 of the IT Act, 2000, operationalised through the IT Rules, 2021 [1] — a framework strong on paper, but uneven in enforcement, as the Delhi High Court's recent direction to Meta to pull down a morphed image illustrates.

Strengths of the existing framework

• Conditional, not absolute, immunity: an intermediary that fails due diligence under Rules 3 and 4 forfeits Section 79 protection [1].
• Time-bound takedown: content exposing private areas, depicting nudity, or impersonating including morphed images must be removed within 24 hours of complaint, which may be filed by the victim or on her behalf [1].
• Victim-facing procedure: MeitY's SOP on NCII (2025) routes complaints through intermediaries, One Stop Centres, police and the National Cybercrime Reporting Portal, with a hash bank for flagged content [2].
• Procedural fairness: the 2025 amendment to Rule 3(1)(d) requires senior-rank, reasoned and periodically reviewed takedown orders [3].
• Synthetic media: the 2026 SGI amendments mandate labelling and traceable metadata for AI-generated content, expressly covering NCII and impersonation [4].

Persisting inadequacies

• Weak enforceability: due diligence carries no penalty; Shreya Singhal (2015) reads "actual knowledge" as a court or government order, so a victim's own complaint does not legally bind the platform [1] — pushing her to litigate.
• Re-upload loop: Rule 4's automated matching targets only content identical to what was removed; a cropped or recoloured version escapes it [1].
• Labelling ≠ consent: a morphed image marked "AI-generated" still destroys reputation [4].
• Investigative gap: police sensitivity and speed in preserving upload trails remain discretionary, not procedural.

The framework is therefore adequate in design but deficient in delivery. Making a completed NCII-SOP complaint count as "actual knowledge", extending hash-based matching to near-identical imagery, and publishing Grievance Appellate Committee outcomes would shift the enforcement burden from the victim to the system — giving substantive content to the constitutional promise of dignity.