On April 7, 2026, the Internet Watch Foundation and Childline published new data from the UK's Report Remove service. The figures are stark: 1,894 reports from children about their own nude or sexual imagery in 2025, a 66% increase in a year. More than a third involved sextortion. The IWF's chief executive described the full scale of the problem as potentially "difficult to fathom." We think platforms need to act on this data, not just observe it.
The numbers
The Report Remove service, run jointly by the IWF and Childline, allows children in the UK to self-report sexual imagery of themselves that has circulated online. What the 2025 data shows is not a gradual drift, it is acceleration.
- 1,894. Reports from under-18s in 2025, up 66% from 2024
- 394. Confirmed sextortion cases, up 34% year on year
- 98%. Of sextortion victims were boys aged 14–17
Of the 1,894 reports, analysts confirmed the presence of child sexual abuse imagery in 1,175, an 83% rise on 2024. Reports involving video climbed 27%, reaching 509 cases. And these are only the children who knew Report Remove existed and felt able to use it. The IWF has been explicit: the real numbers are almost certainly much higher.
On the data ceiling: Report Remove is a UK-specific, opt-in service. Victims must know it exists, trust it, and feel able to come forward. The IWF warns that fear, shame, and stigma mean many cases go unreported entirely. The 394 confirmed sextortion cases are a floor, not a ceiling.
How sextortion actually works
Understanding the harm requires understanding the mechanism. Sextortion against minors is not a single event, it is a process with distinct, detectable stages.
1. Approach and trust-building
Perpetrators contact targets, typically boys aged 14–17, on social platforms or messaging apps. They present as peers, romantic interests, or friendly strangers. This stage often involves love-bombing, flattery, and the establishment of a perceived emotional connection.
2. Image solicitation
The perpetrator requests intimate images or video, often reciprocating with synthetic or stolen imagery to appear credible. This is the point at which detection tools, both on-device and platform-level, have the clearest opportunity to intervene.
3. Escalation and threat
Once imagery is obtained, the tone shifts abruptly. Threats to share the content with the victim's family, school, or social contacts begin. According to the IWF, language escalates rapidly, emotionally manipulative and aggressive. This is the stage most often linked to acute psychological harm and, in the worst cases, suicide.
4. Extortion loop
The victim is coerced into sending money, more images, or both. Compliance does not end the threat. Many cases involve prolonged abuse cycles sustained by the perpetrator's ongoing leverage.
"Criminals are casting their nets wide and are able to corner young people with the most violent and terrifying threats. They employ emotional manipulation and use intimidating, aggressive language and threats that escalate rapidly."
, Kerry Smith, Chief Executive, Internet Watch Foundation
Where AI is making this worse, and where it can help
Generative AI has lowered the operational cost of sextortion at scale. Synthetic imagery makes it easier for perpetrators to appear credible in early grooming stages. AI-assisted scripts allow a single offender to run dozens of simultaneous conversations. And deepfake tools mean that even a child who has never shared an image can be threatened with fabricated material.
The same technology, however, makes earlier detection possible, if platforms build the right infrastructure. The IWF is already calling on government to mandate on-device nudity detection. The OpenAI child protection blueprint published this month makes the case for platform-level synthetic content classification and grooming detection as baseline requirements. The tools exist. The question is deployment.
What Tuteliq can detect, at each stage
Sextortion leaves a detectable signal trail at every stage of the harm lifecycle. Tuteliq's API is designed to catch it across text, image, audio, and video, the full modality surface that modern sextortion campaigns operate across.
--- · --- · ---
Trust-building / grooming · Love-bombing, isolation tactics, secrecy requests, false reciprocity · detect_grooming Multi-turn
Image solicitation · Coercive requests for nude content, synthetic imagery used as bait · detect_unsafe synthetic-content/image
Threat escalation · Aggressive language, blackmail phrasing, escalation patterns in chat · detect_unsafe Critical flag
Sextortion messaging · AI-enhanced extortion scripts, financial demand patterns · synthetic-content → AI_ENHANCED_SEXTORTION tag
Victim distress signals · Fear, shame, isolation indicators in victim's own messages · analyze_emotions
Video / audio evidence · Synthetic or coerced video content, audio transcription for context · synthetic-content/video synthetic-content/audio
A detection call, in practice
Here is what a sextortion detection call looks like when a platform passes a flagged message through Tuteliq's synthetic content endpoint, with age context applied:
POST /api/v1/safety/synthetic-content
{
"text": "I have your photos. Pay me or everyone sees them.",
"context": {
"age_group": "14-17",
"platform": "instagram",
"language": "en"
},
"options": { "include_evidence": true, "support_threshold": "high" }
}
// Response
{
"detected": true,
"level": "critical",
"classification": "confirmed_synthetic",
"categories": [
{ "tag": "AI_ENHANCED_SEXTORTION", "confidence": 0.99 },
{ "tag": "AI_GENERATED_TEXT", "confidence": 0.87 }
],
"age_calibration": { "applied": true, "multiplier": 1.3 },
"recommended_action": "immediate_intervention",
"evidence": [
{ "text": "Pay me or everyone sees them", "tactic": "SEXTORTION", "weight": 0.97 }
],
"support": {
"country": "GB",
"helplines": ["Childline: 0800 1111", "IWF Report Remove"]
}
}Note the support field: when a UK-based victim is identified, the response automatically surfaces Childline and Report Remove, the exact services the IWF is urging more children to use. A platform can pass these directly to a victim in the moment of detection, before the harm escalates further.
This is not a future problem
Nine children a week. That is the current rate in the UK alone, from a service that many victims have never heard of. The IWF has been explicit that what is visible is likely a fraction of what is happening.
The UK government is actively considering mandatory on-device nudity detection. The Online Safety Act is already placing new obligations on platforms. The pressure from attorneys general in the US, from NCMEC, and from the OpenAI blueprint published this month all point in the same direction: baseline detection infrastructure is no longer optional.
Tuteliq is built for exactly this. If your platform handles user-generated content, messaging, or media, and especially if it is accessible to minors, the detection architecture described above can be integrated via API, without rebuilding your trust & safety stack from scratch.
---
Sources: Internet Watch Foundation, April 7 2026 · The Guardian, April 7 2026 · Report Remove / CARE data release. All statistics cited are from the IWF and Report Remove 2025 annual data. Kerry Smith quote is paraphrased from IWF press release.