Remove morphed photo of woman with PM: HC to Meta

1. At a Glance

• Delhi High Court directed Meta to remove a morphed, objectionable image of a woman circulated on social media alongside PM Narendra Modi. [1]
• The case turns on platform (intermediary) liability. Under Section 79 of the IT Act read with the IT Rules, 2021, intermediaries must observe due diligence to keep their safe-harbour exemption. [2]
• Morphed or impersonating intimate imagery attracts a 24-hour takedown duty once a complaint is received. [2]
• Relevant to UPSC because it combines dignity of women, digital governance, judicial oversight of police and online misinformation.

2. Why in the News

• On Friday (25 Sept 2026) the Delhi HC (Justice Girish Kathpalia) ordered Meta to remove the image while hearing the woman's plea. [1]
• The judge called it an "extraordinary position" because it involves the dignity and safety of a woman and also the country's Prime Minister. [1]
• He warned: "If they [Delhi Police] are not sensitive about it, this court will ensure that it becomes sensitive." [1]
• The court issued notice to CJP leaders on the woman's allegation that they helped circulate the image. [1]
• The article does not expand "CJP", so this note does not either.

3. Background & Evolution

• IT Act, 2000: Section 79 gives conditional safe harbour to intermediaries for third-party content. [2][4]
• IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 were notified by the government. [3] Rules 3 and 4 set the due-diligence obligations. [2]
• The 2021 Rules require removal or disabling within 24 hours of complaints about content that exposes private areas, shows nudity or sexual acts, or impersonates, including morphed images. The complaint can be filed by the individual or on their behalf. [2]
• MeitY issued an SOP on Non-Consensual Intimate Imagery (NCII). [5]
• IT Amendment Rules, 2025 revised Rule 3(1)(d) to add transparency, proportionality and accountability safeguards to takedowns. [6]
• MeitY Secretary S. Krishnan has stressed clear, prudent exercise of powers under Sections 69A and 79(3)(b). [7]

4. Core Static Facts

5. Multi-Dimensional Analysis

Legal / Constitutional

• Safe harbour is conditional. An intermediary that fails due diligence loses its Section 79 immunity. [2]
• The court is invoking judicial oversight over both the platform and the police. [1]
• The right to dignity under Art. 21 is engaged. This is my analytical inference; the article does not cite it.

Social

• Morphed imagery of women is a form of gender-based online abuse. The court framed it as a matter of dignity and safety. [1]
• The 2021 Rules let a third party complain on the victim's behalf, which lowers barriers to reporting. [2]

Ethical / Governance

• The court's remark on police sensitivity points to accountability of enforcement agencies. [1]
• The 2025 Rule 3(1)(d) amendment aims to keep takedowns transparent, proportionate and accountable, balancing speech against harm. [6]

Scientific / Technological

• Morphing and synthetic media are a growing platform-governance problem. MeitY's NCII SOP is a procedural response to non-consensual imagery. [5]

Administrative

• Enforcement is split between platforms (grievance redressal), MeitY (rules) and police (investigation). [2][5]

6. Recent Developments (last 12-18 months)

• 2025: MeitY notified amendments to Rule 3(1)(d) of the IT Rules, 2021. [6]
• 2025: MeitY published an SOP on NCII. [5]
• 2025: MeitY Secretary stressed prudent use of Sections 69A and 79(3)(b). [7]
• 25 Sept 2026: Delhi HC directed Meta to remove the morphed image, and notice was issued to CJP leaders. [1]

7. Prelims Hooks

• The Delhi HC directed Meta to remove a morphed image of a woman shown with the PM. [1]
• The bench was Justice Girish Kathpalia. [1]
• Section 79 of the IT Act gives intermediaries conditional exemption from liability for third-party content. [2]
• Due-diligence duties are in Rules 3 and 4 of the IT Rules, 2021. [2]
• Morphed or impersonation intimate content must be removed within 24 hours of complaint. [2]
• A complaint may be filed by the individual or on their behalf. [2]
• The nodal ministry is MeitY. [3]
• Rule 3(1)(d) was amended in 2025 for safeguards on takedowns. [6]
• Section 69A is the blocking power; MeitY has an NCII SOP. [7][5]

8. Why She Needed a High Court Order When a 24-Hour Rule Already Exists

• On paper the platform had to act in 24 hours. In practice she had to reach a judge.
• The IT Rules, 2021 give Meta a 24-hour deadline to pull down a morphed image once a complaint comes in [2].
• But this is a due-diligence duty. If the platform ignores it, no fine follows. The only consequence is that it may later lose safe harbour under Section 79 and be sued [2].

• So the victim carries the cost of enforcement. She must hire a lawyer and move the High Court, as happened here [1].

• The 2015 Shreya Singhal judgment is the reason a private complaint is weak

• The Supreme Court read down Section 79(3)(b). A platform gets "actual knowledge" only from a court order or an order of the government or its agency — not from an ordinary person's email [9].
• That was done to protect speech: a platform should not become a judge of what is legal.

• The side effect is that a woman's own complaint about her own face does not, by itself, legally bind the platform. A judge's word does [1][9].

• The law has a fast lane, but only the State can use it

• Once actual knowledge arrives through a court order or a government notification, removal must happen within three hours [10].
• Three hours for the State, 24 hours for a victim's complaint, and nothing enforceable if the platform simply sits on it — that gap is what the Delhi HC order is filling by hand [1][2][10].

9. Taking Down One Copy Does Not Stop the Next One

• The order removes an image. It does not remove the file.
• A takedown works on a specific post or link. Anyone who already saved the image can upload it again the next minute.

• The court's direction is to Meta [1]. It does not reach WhatsApp forwards, Telegram groups, or any site that is not before the court.

• The Rules do ask for automatic filtering, but the test is too narrow

• Rule 4 of the IT Rules, 2021 requires large platforms (significant social media intermediaries) to use automated tools to spot content that is identical to what was already removed [2].
• "Identical" is the weak word. Crop the picture, add a caption, change the colour, or re-morph the same face, and the match fails.

• So each new version has to be reported afresh, and the 24-hour clock starts again [2].

• This is why the judge turned to the police, not only to Meta

• Removing copies is endless. Finding and charging the person who made and spread the image is what actually stops it.

• That job sits with Delhi Police, not with MeitY or the platform [2][5] — which is why the court warned that if the police are not sensitive, "this court will ensure that it becomes sensitive" [1].

• The court also went after the spreaders, not just the host — notice was issued to CJP leaders on the allegation that they helped circulate the image [1]. That is a shift from content control to liability of the people forwarding it.

10. The Case Against Quick Takedowns — and Why It Does Not Win Here

• The strongest opposing argument is real: fast takedowns get misused.
• If a platform loses safe harbour for guessing wrong, the safe choice is always to delete. Over time it deletes first and asks later.
• Genuine criticism, satire and journalism get caught in that net. Nobody is compensated for a wrongly deleted post.

• The government itself accepted this risk. The 2025 amendment to Rule 3(1)(d) added transparency, proportionality and accountability safeguards to takedown orders [6]. MeitY Secretary S. Krishnan asked for prudent and clear use of Sections 69A and 79(3)(b) [7].

• Concede what is right in it — a one-sided removal power, used on political content, is censorship without a hearing. That is exactly why Shreya Singhal insisted on a court or government order [9].

• But the argument does not carry over to this case
• A morphed intimate image of a named woman has no version that is lawful. There is no public interest to weigh on the other side.
• The harm is one-way. A wrongly removed article can be restored; an image seen by thousands cannot be unseen.
• So the honest answer is to split the categories: keep slow, reasoned, reviewable process for speech about public affairs, and keep the fast lane only for non-consensual and morphed intimate imagery, which the Rules already treat as a separate class with its own 24-hour clock [2][5].

11. The New Deepfake Rules Help the Viewer, Not This Victim

• What changed in 2026 — MeitY notified amendment rules on synthetic generated information (SGI), that is, audio or video made or altered by a computer so well that it looks real. They came into force on 20 February 2026 [8].
• Their logic is labelling, and labelling solves a different problem
• The aim is to stop viewers being deceived — so synthetic content must be marked as synthetic [8].
• A woman's morphed photo is not mainly a deception problem. Even a picture clearly stamped "AI-generated" still destroys her reputation and safety.

• Consent, not honesty about the tool, is the missing element. A label does not restore consent.

• What is still useful in it for her — a duty to mark and trace synthetic content gives police and platforms a technical trail to follow when asking who generated the file [8], which is the part the Delhi Police must now do [1].

12. What Would Stop the Next Victim From Needing a Judge

• MeitY should widen Rule 4's automatic matching from "identical" to "near-identical" for intimate imagery
• Today the duty only catches exact copies of removed content [2].

• Matching by image fingerprint, so a cropped or recoloured version of the same picture is also blocked, would stop the re-upload loop without any new complaint from the victim.

• MeitY should make the NCII SOP the single door, and make its outcome binding

• The SOP on Non-Consensual Intimate Imagery already lays out a victim-facing procedure [5].

• If a completed SOP complaint counted as "actual knowledge" the way a government notification does under Section 79(3)(b), the three-hour lane would open to the victim instead of only to the State [5][10].

• Use the Grievance Appellate Committee route, and publish its record

• The 2021 Rules created Grievance Appellate Committees so a user whose complaint is ignored by a platform has a free appeal, without going to court [2].

• Publishing how many NCII appeals each platform lost would show whether the 24-hour duty is being met at all — at present nobody outside the platform knows.

• Delhi Police should treat morphed-image complaints as time-bound

• The judge's warning was about sensitivity, not about law [1]. Sensitivity is not a procedure.
• A fixed internal timeline for preservation requests to the platform, before the account is deleted and the upload trail is lost, is the concrete version of what the court asked for.

13. Anchors for Answers

• Data: 24 hours to remove morphed or intimate imagery on complaint; 3 hours to remove once a court or government order arrives [2][10]
• Law/Case: Section 79 and Section 79(3)(b), IT Act, 2000; Section 69A; Rules 3, 4 and Rule 3(1)(d) as amended in 2025, IT Rules, 2021 [2][6][7]
• Law/Case: Shreya Singhal v. Union of India (2015) — "actual knowledge" means a court or government order, not a private complaint [9]
• Law/Case: Right to dignity and privacy under Article 21 (K.S. Puttaswamy) — my analytical framing, not from the article
• Scheme: MeitY SOP on NCII — the victim-facing takedown procedure [5]; Grievance Appellate Committees under the 2021 Rules as the free appeal against platform inaction [2]
• Recent rule: IT (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026 on synthetic generated information (deepfakes), in force 20 February 2026 [8]

14. Mains Relevance

• GS-II: Government policies and interventions; role of the judiciary; issues relating to women.
• GS-III: Cyber security; role of media and social networking sites; internal security challenges.
• GS-IV: Ethics in governance and accountability.
• Plausible questions:
• "Safe harbour under Section 79 is conditional, not absolute." Discuss with reference to morphed content and platform accountability.
• Examine the adequacy of India's legal framework against non-consensual and morphed imagery of women.
• How can courts and police be made more responsive to online gender abuse without curbing free speech?

15. Related Topics to Study Next

• Shreya Singhal v. Union of India (2015): reading down of Section 79 and striking down of Section 66A. Study the Section 79 "actual knowledge" standard.
• Section 69A blocking and its procedural safeguards, since the state's direct takedown power sits alongside Section 79.
• NCII SOP of MeitY, the victim-facing procedure. [5]
• Deepfakes / synthetic media regulation, the technology behind morphing.
• Right to privacy (Puttaswamy) and dignity under Art. 21.
• Grievance Appellate Committees under the 2021 Rules.
• Digital Personal Data Protection Act, 2023, for data and consent.
• Personality rights litigation in Delhi HC, the same court's approach to misuse of images.

16. Common Errors / Trap Areas

• Section 79 is safe harbour. Section 69A is the blocking power. Do not confuse them. [7]
• The 24-hour limit applies to intimate, impersonation and morphed content complaints, not to all content. [2]
• The IT Rules, 2021 are subordinate legislation under the IT Act, not a separate Act.
• The Rules are administered by MeitY, not MIB or MHA. MIB has a role in digital media ethics. This distinction is my own gloss.
• Do not assume "CJP" means a specific party. The article does not expand it. [1]