DHS · DHS-362 · record R1913

Facial Recognition for Investigations of Child Sexual Exploitation and Abuse

Department of Homeland Security · ICE

Overview

Development stage: DeployedSource: c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
High-impact designation: High-impactSource: a) High-impact
Impact justification: Not reported
Start date: 2020-12-01 (day precision)Source: 2020-12-01T00:00:00
Withheld from public reporting: Not reported

Mission

Topic area: Law Enforcement
Operational functions: Biometrics and identity Derived by keyword rules; see methodology

Problem the AI is intended to solve

This use case intends to solve the problem of identifying unknown victims and offenders depicted in child sexual abuse material.

Expected benefits

The tool helps more quickly identify previously unknown victims and offenders who might not be discovered through manual investigative methods alone. By highlighting potentially relevant photographs or profiles across publicly available online images, it can accelerate victim identification and rescue efforts and support the disruption and prosecution of offenders who might otherwise remain undetected.

System outputs

Homeland Security Investigations Child Exploitation Investigations Unit personnel submit newly discovered and unidentified child sexual abuse material images obtained through lawful investigative means to an AI-enabled facial recognition service. The service compares these photos to publicly available online images to find visually similar faces. It returns possible matches, along with links or references to the public sources where those images were found, so investigators can review them in context. These results are treated as investigative leads that may point to potential identities or locations, but they do not constitute confirmation on their own.

Technology

AI classification: Computer visionSource: Computer Vision: AI that processes and interprets visual data (e.g., images and videos).
System name(s): Law Enforcement Sensitive (LES)
Custom-developed code: No
Public source code: Not reported

Sourcing

How it was built: Purchased from vendorSource: a) Purchased from a vendor
Vendor (as reported): Law Enforcement Sensitive (LES)
Vendors (standardized): None on the standard list

A vendor is the supplier named by the agency. It does not identify the underlying model or AI technology.

Data and privacy

Involves PII: Yes
Privacy Impact Assessment: www.dhs.gov ↗
Authorization to Operate: Yes
Demographic features: Not reported
Training and evaluation data: Law Enforcement Sensitive (LES)
Federal Data Catalog: Not reported

Governance

8 of 8 minimum-practice questions answered. A blank answer means the agency reported nothing; it does not mean the practice is absent.

Pre-deployment testingReported in placeSource: a) Yes

AI impact assessmentReported in placeSource: a) Yes

Independent reviewReported in placeSource: c) Yes – by the CAIO

Ongoing monitoringReported in placeSource: a) Yes, sufficient monitoring protocols have been established

Operator trainingReported in placeSource: a) Yes, sufficient and periodic training has been established

Fail-safeNot applicableSource: b) Not applicable

Appeal processNot applicableSource: b) Not applicable

User and public consultationOther or ambiguous answerSource: d) Other

Potential impacts and how they were identified

The AI-enabled facial recognition service may return too many candidates, resulting in the collection of irrelevant personal information. Mitigation: The service only returns candidates meeting a set confidence score threshold, ranking results by highest confidence. Potential matches are used as investigative leads and require full validation through the investigative process.