AETERNAE AI RESEARCH LLC INDEPENDENT RESEARCH
ÆAETERNAERESEARCH
Sign inRequest access
← Malware catalog

Attributed metadata history

REPORTPUBLIC METADATA

ICE seeks voice biometrics to identify detainees, outside callers in AI surveillance system

U.S. Immigration and Customs Enforcement (ICE) is seeking technology to turn recorded detainee calls and live surveillance video into a searchable intelligence network capable of identifying speakers by their voices, tracking movements, and mapping relationships across detention facilities. The proposed AI system would build voice profiles without subjects’ cooperation, analyze conversations and behavior, and automatically connect findings to investigative records to reconstruct individuals’ activities and associations. The system would also analyze detainees’ behavior through continuous video surveillance, identify associations among individuals, detect potentially suspicious conversations, and automatically correlate information from otherwise separate detention and law enforcement systems to build investigative timelines and evidence packages. The proposed capability also could extend beyond people in ICE custody. The unusually detailed requirements are contained in a 73-page Performance Work Statement accompanying a Request for Information (RFI) that ICE published last week. The agency is seeking industry capabilities for what it calls an Integrated Intelligence Suite, designed for deployment across detention facilities operated by ICE, private contractors, and government partners. Among its most consequential requirements is the ability to create biometric voice profiles from existing recordings without a separate enrollment session, subject cooperation, or new recording equipment. Investigators could then search for potential matches regardless of the telephone number, account, personal identification number (PIN), or device used. Where available audio and government authorization permit, the system would compare the voice of someone receiving a detainee's call against a profile associated with the dialed number, potentially revealing call forwarding or other arrangements that obscure who is participating in a conversation. Investigators could also submit an approved recording obtained from another source, such as an interview or investigative tip, and search the authorized detention communications database for matching speakers. Where separately authorized, the system would support matching voices across participating agencies. The proposed system would not be limited to identifying speakers. ICE wants to combine potential voice matches with video footage, detainee-management records, housing assignments, movement histories, and incident reports, allowing investigators to identify relationships and reconstruct activity across multiple facilities. The procurement remains at the market research stage. ICE has not identified a selected contractor, contract value, deployment schedule, or the number of facilities that would ultimately use the system. Vendor responses are due October 26. A single system for reconstructing activity and associations Perhaps the most far-reaching element of the proposal is its requirement for automated correlation across different kinds of records. The system would extract and link names, aliases, telephone numbers, organizations, locations, communications, video events, and investigative information. ICE wants those connections represented as searchable relationships and timelines, with the evidence supporting each asserted connection available to investigators. The agency’s mandatory acceptance test illustrates what it expects. Starting with a flagged detainee communication, the system must present the recording, transcript, translation (when applicable), potential speakers, and any discrepancies in accounts. It must then retrieve relevant detainee records, identify related video events, and display connections among the people, locations, and events involved. The investigator could assemble a case record and export an evidence package containing supporting audio, video, metadata, relationships, and an audit trail without manually moving information between unrelated a

Reference details

Source category
research-report
First reported
Oct 13, 2026 19:49 UTC
Last reported
Oct 13, 2026 19:49 UTC
Catalog updated
Oct 14, 2026 04:13 UTC
Record ID
decc4d58ea305d826fca445146c75ca29c3d1172900170076ae09be6fca1506d

Published indicators

Network indicators are visibly defanged. This page does not make them clickable or send them to another service.

No validated file hashes or network indicators are stored for this record.

Source observations

intelligence_src_df3440d5202f514d9e

Source-provided label
Publisher RSS metadata
Evidence class
mentioned in report
Confidence
Not reported
Observed
Oct 13, 2026 19:49 UTC
Retrieved
Oct 14, 2026 04:13 UTC

This record contains metadata and attribution only. It does not host, download, unpack, render, or execute malware samples.

Feed collection status

Sources represented on this page. Publication and observation dates describe the content; successful batch dates describe local collection. A recent batch does not certify complete coverage or current provider availability.

1 sources: 0 stale, 0 needing attention, 0 paused, 0 unverified.

Review source update dates and status
  • intelligence_src_df3440d5202f514d9eRecent successful batchLast successful batch: Last attempt:

Stale means older than twice the recorded collection interval, with a minimum of one hour; without an interval, the threshold is 24 hours. Cached content remains available while a feed is stale, paused, or unverified.