Skip to main content
Incident intelligence/SS-IR-035CASE FILE OPEN
Symbolic editorial illustration for SS-IR-035SERVANTSTACK // INCIDENT INTELLIGENCEFORENSIC IMAGE // VERIFIED FRAME
SS-IR-035 // INCIDENT REPORTOfficial finding

ClothOff

An AI Nudify App Manufactured Child Abuse Images of 20+ Schoolgirls in One Spanish Town

EXECUTIVE BRIEF

In September 2023, in the town of Almendralejo, Spain, more than 20 girls aged 11 to 17 discovered that fake nude images of themselves were circulating in local WhatsApp groups.

DEEPFAKES & FRAUDSAFETY & SOCIETY
FAILURE CHAINTRACE COMPLETE
  1. 01TRIGGERIn September 2023, in the town of Almendralejo, Spain, more than 20 girls aged 11 to 17 discovered that fake nude…
  2. 02MACHINE ACTIONAutonomous actor
  3. 03MISSING GATEIdentity verification and dual control
  4. 04IMPACTHuman welfare
01 // INCIDENT SUMMARY

The short version

In September 2023, in the town of Almendralejo, Spain, more than 20 girls aged 11 to 17 discovered that fake nude images of themselves were circulating in local WhatsApp groups.

02 // KEY FACTS

Case telemetry

INCIDENT
SS-IR-035
DATE
September 2023
SYSTEM
ClothOff
LOCATION / SCOPE
Almendralejo, Extremadura, Spain
EVIDENCE
Official finding
AI ROLE
Autonomous actor
HARM
Human welfare
SOURCES
2 cited records
03ENTRY POINT // WHAT HAPPENED

The event

In September 2023, in the town of Almendralejo, Spain, more than 20 girls aged 11 to 17 discovered that fake nude images of themselves were circulating in local WhatsApp groups. A group of boys, many of them classmates, had pulled the girls' clothed photos from their Instagram profiles and fed them to ClothOff, an AI "nudify" app that advertised "Undress anybody" and charged roughly 10 euros to generate 25 fabricated nude images. At least one victim reported being blackmailed with a fake image. A youth court in Badajoz later convicted 15 minors on 20 counts of producing child abuse material and 20 counts against the victims' moral integrity, sentencing them to one year of probation. The case became a landmark for nudify-app harm and exposed that Spanish law had no clean statute for AI-generated intimate images of minors.

04CAUSAL TRACE // AI'S ACTUAL ROLE

What the machine did

ClothOff is a single-purpose generative model built to do exactly one thing: take any clothed photo and synthesize a realistic nude body underneath the real face. There was no consent check, no age check, no human in the loop, and no validation that the subject was an adult or had agreed. A child's face uploaded by a stranger was processed identically to any other input. The model ran fully autonomously at the speed and scale of a paid API call, producing photorealistic child sexual abuse material in seconds with zero oversight gate between the upload and the output. The "feature" was the harm.

Autonomous actorAutomation was a causal participant—not a decorative label for the system around it.
05BLAST RADIUS // CONSEQUENCES

Where the failure landed

More than 20 minors were victimized, the youngest only 11 years old. Fabricated nude images of named, identifiable children spread through their own community on WhatsApp; at least one girl was extorted. Fifteen minors were criminally convicted and placed on a year of probation. The episode triggered national outrage in Spain, drove EU and global policy debate over nudify apps and AI-generated CSAM, and revealed a legal gap that prosecutors had to stretch existing child-abuse-material and moral-integrity statutes to fill. The reputational and psychological damage to the victims is permanent; the images, once distributed, cannot be fully recalled.

06 // EVIDENCE STATUS

Official finding

Supported by a court, regulator, inquiry, or other official record cited below.

SOURCE RECORD UPDATED 2026-07-09

07 // SOURCE LEDGER

2 cited records

  1. 01
  2. 02
08CONTROL FAILURE // MISSING GOVERNANCE

Identity verification and dual control

The failure pattern in this case: Unverified identity or synthetic media.

09INTERVENTION POINT // HUMAN IN THE MIDDLE

The moment the path could change

A named reviewer verifies identity through a separate trusted channel before money, access, or public claims can move.

AI PROPOSES→HUMAN OWNS THE DECISION→SYSTEM EXECUTES
10CONTROL DEPLOYMENT // AUTHORITYGATE

Identity verification · dual control

AuthorityGate's Operational Resilience framework treats any image-transformation pipeline that produces intimate or biometric output as a high-risk action that cannot complete without passing a mandatory pre-generation validation gate. A human SME review gate would be required to define and enforce non-negotiable input controls before a single image is ever processed: verified consent from the depicted subject, an age-assurance check that hard-blocks any input assessed as a minor, and provenance validation rejecting scraped third-party social media images. Any model whose intended output is a synthetic nude is flagged at the change-validation gate as a prohibited use case and never reaches production. There is no autonomous path from "stranger uploads a child's Instagram photo" to "photorealistic nude image returned." The gate fails closed: no consent, no verified adult age, no provenance, no output.

RELEVANT KEYSTONE CONTROLHuman-in-the-Loop ValidationHow high-risk actions route to a named subject-matter expert who owns the go or no-go decision.
12 // THE ALTERNATIVE

Autonomy is a design choice.

See the operating model that keeps AI useful while preserving human authority at consequential moments.

Compare AgenticAI and AugmentedAI →