India digital gender-based violence

India's New Three-Hour Deepfake Takedown Clock Fixes Platform Speed, Not the Cross-Border Enforcement Gap Sinking Korea's Crackdown

Seoul's 2025 crackdown shows criminalizing deepfakes fails without foreign-platform cooperation — India's new takedown rule shares the same blind spot.

Deepfake Sex Crimes: The Enforcement Gap People of Internet Research · India 35.2% Of Korea's cybersex crimes 1,553 deepfake offenses were the l… 42% Foreign platform requests unanswered Only ~58% of 26,697 data requests … 47.6% Suspects who are teenagers 1,761 of 3,557 people apprehended … 3 hours India's new takedown deadline MeitY's Feb 2026 IT Rules amendmen… peopleofinternet.com
Deepfake Sex Crimes: The Enforcement G… People of Internet Research · India 35.2% Of Korea's cybersex crimes 42% Foreign platform requests unanswe… 47.6% Suspects who are teenagers 3 hours India's new takedown deadline peopleofinternet.com

Key Takeaways

The number that should worry Delhi, not just Seoul

At an Amnesty International Korea forum on August 27, 2026, police data laid bare how badly law enforcement is losing the fight against deepfake sexual abuse. Deepfake offenses made up 1,553 cases — 35.2% of all cybersex crimes — in South Korea's 2025 crackdown. Of the 3,557 people apprehended, 1,761, or 47.6%, were teenagers. And when Korean police sent 26,697 data requests to foreign platforms to identify suspects, only about 58% got a response (roughly 15,466 requests answered) (Korea Times). Korea has criminalized deepfake sex crimes since 2020. It still can't out-run the problem, because the bottleneck was never domestic law — it's what happens when the evidence sits on a server in another jurisdiction.

India is walking into the identical trap, just with a different fix.

India's mirror problem

The National Commission for Women has been sounding this alarm since well before 2026. Cybercrime cases against women climbed from 52,974 in 2021 to 65,893 in 2022, prompting the NCW to open a dedicated Digital Shakti Center in New Delhi for cyber-harassment complaints (NCW/Akashvani). By November 2025, the trajectory had worsened enough that the NCW formally asked the Ministry of Home Affairs, MeitY, and the Ministry of Women and Child Development to redefine "deepfake abuse" as a distinct offense — arguing that India's laws still route synthetic-media harassment through analogue-era provisions on voyeurism and defamation, with no dedicated statutory language for non-consensual synthetic intimate imagery (The Print).

What the government built instead

MeitY's answer arrived three months later. On February 10, 2026, it notified amendments to the IT Rules that define "synthetically generated information" for the first time and force platforms to act fast: government or court-flagged deepfake content must come down within three hours, non-consensual intimate imagery within two, and permitted AI content must carry a tamper-resistant label. The rules took effect February 20, 2026, after a ten-day compliance window (Akashvani/Government of India).

Steelmanning the speed obsession

There is a real case for a fast clock. Non-consensual intimate imagery compounds with every hour it stays live — it gets screenshotted, re-uploaded, and mirrored across platforms the original takedown order never reached. When roughly half the suspects in a country's biggest deepfake crackdown are minors, as Korea's data shows, speed genuinely matters both to limit a victim's re-traumatization and to stop peer-to-peer circulation among teenagers before it becomes normalized. A three-hour statutory deadline is a defensible response to a genuinely viral harm, and India was right to shorten a 36-hour window that was clearly too slow for this category of content.

But the Korea data points at a different chokepoint

What a domestic takedown clock cannot touch is the number that actually broke Korea's crackdown: a 42% non-response rate on data requests to the foreign platforms hosting the content. India's deepfake problem runs through the same handful of global platforms — Meta, X, Telegram, Google — via the same slow, largely voluntary channels for law-enforcement data requests. Compressing a platform's removal deadline to three hours does nothing to compress the weeks or months it can take that same platform's trust-and-safety or legal team, sitting in Dublin or California, to respond to an Indian police request for subscriber data needed to actually prosecute. Korea criminalized the conduct years ago and still hit this wall. India is now regulating platform content without a parallel push on platform cooperation with investigators — the actual gap the Seoul forum was convened to discuss.

A two-track reality already visible in Indian courts

The unevenness shows up in who gets fast relief. Delhi High Court has granted ex-parte injunctions within days for well-resourced plaintiffs — ordering platforms to pull deepfakes and disclose account data for figures like a Shark Tank India judge and a sitting MP — because they could litigate directly. An ordinary complainant instead files through cybercrime.gov.in and waits on the same police pipeline the NCW says is still operating without a specific legal definition of the harm. The three-hour rule formally applies to both, but only one track has a lawyer accelerating it.

The proportionate fix

The evidence-based response isn't to abandon takedown speed — it's to stop treating it as the whole answer. India should pair the IT Rules' labeling and removal timelines with the specific criminal provision the NCW has already asked for, so prosecutors aren't stretching voyeurism and defamation clauses to fit synthetic media. And it should push, through the same diplomatic channels used for tax and data-localization disputes, for enforceable response-time commitments from platforms on law-enforcement data requests — the exact metric Korea's forum showed is failing at a 42% rate. A rule that only makes content disappear faster, without also making perpetrators easier to identify and charge, treats the symptom Seoul is watching spiral and leaves the underlying disease alone.

Sources & Citations

  1. Korea Times: deepfake sex crimes surge
  2. Govt sets 3-hour deepfake takedown deadline
  3. NCW inaugurates Digital Shakti Center
  4. The Print: NCW seeks review of deepfake laws
  5. Mondaq: IT Rules 2026 deepfake regulation analysis