What the numbers show
On 27 September 2026 the National Office of Investigation at the Korean National Police Agency reported 598 undercover investigations into digital sex crimes between January and August 2026, against 179 in the same months of 2025. Yonhap, carried by The Korea Times, says that is more than triple the earlier figure. Since the tactic began in 2021, police have run 1,503 such investigations through the end of August, apprehended 2,873 suspects and formally arrested 164.
The same report says the government extended undercover work to cases with adult victims in June 2025. It also says a law criminalizing attempts to start conversations with minors for sexual exploitation took effect in October 2025. It describes the cases as often involving deepfake technology. In the child-focused stings, 338 suspects were apprehended from January to August 2026 and six were formally arrested.
The strongest case for the stings
The case for this tactic is serious, and a policy publication should state it plainly. The programme grew out of the "Nth room" case, in which girls were coerced into abusive acts on camera and the footage was sold in paid chatrooms. Those spaces are closed, encrypted and invite-only. Victims often cannot report, and ordinary complaint-driven policing never sees them. An officer posing as a participant may be the only way in. Korea's own digital sex-crime caseload shows the scale: according to a Yonhap report in The Korea Times, police investigated 17,716 digital sex-crime cases from 2021 to 2025, and the number of victims receiving support reached 10,637 last year, up from 2,087 in 2019.
The legal foundations are also not improvised. The government's April 2020 package of measures to eradicate digital sex crimes planned to let investigators pose as minors and to extend techniques previously reserved for drug cases, with a legal basis to be created. It defined digital sexual abuse material to include synthesized or edited content such as deepfakes. The statute that covers false video products is listed in the Korea Law Translation Center's English edition of the Act on Special Cases Concerning the Punishment of Sexual Crimes as Article 14-2. The listing I could retrieve shows article titles only, so this analysis does not rely on its penalty terms.
Why activity is not the same as effectiveness
A threefold rise in operations is evidence that police are doing more. It is not evidence that more harm is being prevented. The published figures stop at apprehension. The 164 formal arrests are about 5.7 percent of the 2,873 suspects apprehended since 2021. In the 2026 child-focused stings, 6 of 338 suspects, about 1.8 percent, were formally arrested. A low arrest rate may reflect cautious use of detention, since arrest and detention are the most intrusive steps. It may also reflect weak cases, or suspects who cannot be traced. The public cannot tell which, because the data does not report indictments, convictions, dismissals or acquittals.
That gap matters because undercover work carries a risk that ordinary policing does not. When officers pose as minors or as participants, the line between investigating a crime and inducing one depends on supervision and records. The 2025 expansion to adult victims and the October 2025 offence of attempted grooming both reach conduct at an early, preparatory stage. An offence defined by attempt makes the quality of the police-created conversation central to guilt. Courts and defence lawyers can test that only if the practices are documented.
Scope creep is the real regulatory risk
The stings are one part of a wider package. In May 2025 the gender ministry announced its 5th Comprehensive Measures to Protect Youth (2025 to 2027). They include requiring platforms to preemptively block suspected sexual exploitation content, even where its legality is unclear, and to ask the Korea Communications Standards Commission to review it. In August 2026, according to the Yonhap report, the government said it would review allowing investigators to lawfully hack websites that distribute illegally filmed material. It also floated criminalizing the opening of such sites and notifying foreign authorities about overseas sites that fail to remove material involving Korean victims. The lawful-hacking idea is still at the review stage; no bill or enactment date was given.
Our view is that targeting the production and distribution of non-consensual sexual imagery is legitimate and urgent. Blocking content whose legality is unclear is a different matter. Platforms facing liability for under-blocking will over-block, and the cost falls on lawful speech and on the survivors' own advocacy and reporting. The Korean government itself noted the problem at scale: the report cites 34,628 identified websites with illegally filmed material, 6,683 still active. A takedown-by-default regime across tens of thousands of sites invites errors with no appeal path.
What proportionate oversight would look like
None of this argues for scaling back undercover investigation of child exploitation. It argues for the accountability that makes such powers sustainable. Four steps would be proportionate and cheap:
- Publish the funnel. Report apprehensions, referrals, indictments, convictions and dismissals each year, split by child-victim, adult-victim and deepfake cases.
- Document inducement safeguards. Publish the rules on when an officer may initiate contact and how conversations are logged for court review.
- Require judicial authorization for hacking. The Korea Times report notes that similar warrant-based systems exist in Australia and Britain, and any Korean power should be warrant-based with notification and review.
- Add a sunset and an appeal route for any preemptive-blocking duty, so wrongly blocked material can be restored quickly.
South Korea has built one of the world's most active enforcement regimes against digital sexual abuse, and the stings are a defensible part of it. The next test is whether the state will show, with data, that the growth in operations converts into convictions and fewer victims, rather than asking for trust on the strength of activity counts.