On September 30, 2026, a Delhi High Court division bench of Justices C. Hari Shankar and Nivedita Anil Sharma clarified that the de-indexing directions in the court's May 29 right-to-be-forgotten judgment apply to search engines and legal databases, not to Meta's platforms such as Facebook, according to MediaNama's report. The ruling is narrow. It still shows how a judge-made privacy remedy gets its boundaries, and how much work Indian law has left undone.
What the May 29 judgment did
In Laksh Vir Singh Yadav v. Union of India, Justice Sachin Datta decided a consolidated batch of more than 30 petitions. Many came from acquitted or discharged people and from parties to resolved matrimonial disputes. LiveLaw's summary says the court grounded the right in informational privacy under Article 21. It separated de-indexing, which hides a page from name-based search results while the source stays online, from masking, which replaces names in digital copies of judgments with neutral references. According to IndiaLaw's analysis, the judgment also set a seven-factor test that includes the passage of time, the person's public role and the effect on freedom of expression. It barred relief entirely for convictions involving offences against women or children and for breaches of public trust. The same analysis reports a two-week deadline for search engines and a four-week compliance affidavit from MeitY.
The constitutional anchor is the Supreme Court's nine-judge decision in Justice K.S. Puttaswamy (Retd.) v. Union of India (Writ Petition (Civil) No. 494 of 2012, decided August 24, 2017). That decision recognised privacy as a fundamental right.
What the September 30 clarification says
Meta asked the court whether the directions reached Facebook. Senior Advocate Mukul Rohatgi told the bench that Meta had no grievance with the judgment and had already taken down the URLs in question. In his words, "I am Facebook, I am not a search engine." MediaNama reports that the opposing parties did not object, the bench agreed, and Meta's plea was closed. Indian Kanoon's appeals are separate. Eight are pending, five before a division bench. The remaining three are listed for October 6, when the court will examine whether restricting name-based search adequately serves lawyers' professional needs.
The case for a broader reading
The strongest argument against Meta's position is practical. For an acquitted person, a Facebook post or group that repeats an old court record can be as damaging as a search result. Privacy harms follow the content, not the product category. A court that cares about dignity might therefore want one rule for every intermediary that amplifies the record.
The argument has a real cost, though. The remedy in the May judgment is name-based de-indexing. It works because a search engine builds a retrievable index keyed to names. A social network hosting a post is a different function. The right tool there is takedown of specific content, and Meta says it has already done that for the URLs complained of. Extending an indexing remedy to a platform with no index would give the order nothing to bite on. It would also invite over-removal as a way to avoid contempt risk.
Why functional scoping is the right instinct
The division bench's clarification applies a principle that is good for both speech and innovation: obligations should track the function a service performs. Europe's template, Google Spain (CJEU, C-131/12, cited in the May judgment per IndiaLaw), also focused on search engine operators. Delisting there is not blanket erasure, and the source page stays online.
Proportionality is what keeps the remedy compatible with open-court principles. Under this framework the judgments stay accessible by case number, citation, court and date. The record is not deleted. Only the ability to type a person's name and surface decades-old proceedings is limited. The exclusions for serious offences and breaches of public trust also protect the public's legitimate interest in accountability.
Some risks remain. The Indian Kanoon appeals will test whether limiting name search harms lawyers, researchers and journalists who rely on it. Smaller legal-information services lack Google's compliance budget. A two-week de-indexing deadline across thousands of requests could push platforms towards removing first and reviewing later.
The statutory gap
The deeper problem is that this regime exists only because courts filled a vacuum. Section 12 of the Digital Personal Data Protection Bill, 2023 as introduced is headed "Right to correction and erasure of personal data." It is a consent-based erasure right, not a general right to be forgotten. The text as introduced also excluded from its scope personal data that someone is legally obliged to make public. IndiaLaw notes that the May judgment leaves the relationship between the DPDP Act's erasure right and judicial-records cases "not definitively resolved." Statutory silence leaves search engines, legal databases and social platforms guessing which duties apply to them.
That uncertainty is the real cost for innovators. Each new case, like Meta's plea here, becomes a miniature rulemaking exercise.
What should happen next
- Keep the remedy narrow. Name-based de-indexing of judicial records, with the source preserved, is a proportionate tool. The Delhi HC's scoping keeps it that way.
- Protect lawful research access. The October 6 hearing should test whether access by case number, court and date really serves practitioners and journalists.
- Legislate and write procedure. MeitY and Parliament should state how erasure under the DPDP framework interacts with court records, with clear timelines, an appeal route and transparency reporting. Courts should not have to supply these through case-by-case orders.
The September 30 order is a small win for regulatory clarity. It matches obligation to function and leaves broader questions for the appeals and, eventually, for lawmakers.