60 lines
4.1 KiB
JSON
60 lines
4.1 KiB
JSON
{
|
|
"title": "Alvin, Simon, and... Al Qaeda? Finding Modified Voice in SIGINT Traffic",
|
|
"type": "SIDTODAY",
|
|
"description": "This SIDtoday article discusses the use of voice modification technology by Al-Qaeda and other extremist groups to evade identification in SIGINT traffic. It highlights the development of the HLT Lite software by R64, which can automatically detect modified voices in intercepted communications, enabling the identification of targets even when traditional voice recognition fails.",
|
|
"long_description": "This is a TOP SECRET//SI//REL TO USA, FVEY SIDToday article dated 2 June 2010, authored by a researcher in the Human Language Technology Research Office (R64). The article describes the discovery of voice modification technology in SIGINT collection and the development of an automated detection capability in response.\n\nThe operational trigger was the discovery in late 2009 of two audio cuts associated with Al Qaeda in the Arabian Peninsula (AQAP) in Yemen containing clearly modified voice — described as sounding like a character from Alvin and the Chipmunks. R64 and CES (S31) were tasked to recover the original voice for speaker identification purposes. Within two weeks a profile of the modification technique used by the AQAP handset had been created, and CES deployed a reversal capability available internally at \"go saas.\"\n\nThe follow-on research question was whether modified voice could be automatically detected in bulk SIGINT traffic for target discovery purposes. This became operationally urgent after the December 25 2009 underwear bombing attempt in Detroit, when many AQAP selectors in Yemen went quiet — raising the question of whether targets had switched to new phone numbers while continuing to use voice modification. In response R64 developed HLT Lite, a software package using three core algorithms: speech activity detection, anomaly detection and automatic speaker recognition. HLT Lite was deployed to the Special Collection Service site in Yemen on 9 March 2010 to search JUGGERNAUT collection data. Within 30 minutes of deployment it had discovered a cut containing modified voice. Over the following period it scanned more than one million audio cuts and found at least 80 confirmed examples, leading to the tasking of four phone numbers with more under analysis.\n\nA broader search of 500,000 NUCLEON cuts found 76 additional examples of modified voice across multiple product lines including Afghanistan/Taliban, Sunni Extremism, Lashkar-e Tayyiba and Southwest Asia Narcotics — establishing that voice modification was not limited to AQAP but was a cross-target phenomenon.",
|
|
"tags": [
|
|
"R64",
|
|
"voice modification",
|
|
"HLT Lite",
|
|
"Al-Qaeda",
|
|
"AQAP",
|
|
"Human Language Technology",
|
|
"anomaly detection",
|
|
"speech recognition",
|
|
"target discovery",
|
|
"counterterrorism",
|
|
"modified coice",
|
|
"Alvin and the Chipmunks"
|
|
],
|
|
"document": {
|
|
"creator": "NSA",
|
|
"date_created": "2010-06-02",
|
|
"classification": [
|
|
"TOP SECRET",
|
|
"SI"
|
|
],
|
|
"dissemination": [
|
|
"REL"
|
|
],
|
|
"banner": "DYNAMIC PAGE -- HIGHEST POSSIBLE CLASSIFICATION IS TOP SECRET // SI / TK // REL TO USA AUS CAN GBR NZL",
|
|
"genre": "newsletter",
|
|
"programs": [
|
|
"JUGGERNAUT",
|
|
"HLT",
|
|
"NUCLEON"
|
|
],
|
|
"countries": [
|
|
"Yemen",
|
|
"Afghanistan",
|
|
"Iraq"
|
|
],
|
|
"people": []
|
|
},
|
|
"publication": [
|
|
{
|
|
"publisher": "The Intercept",
|
|
"date_published": "2018-01-19",
|
|
"title": "Finding Your Voice: Forget About Siri and Alexa — When It Comes to Voice Identification, the “NSA Reigns Supreme”",
|
|
"link": "https://theintercept.com/2018/01/19/voice-recognition-technology-nsa/",
|
|
"reporters": [
|
|
"Kofman, Ava"
|
|
],
|
|
"contributors": [],
|
|
"doc_source": "https://theintercept.com/document/2018/01/19/finding-a-modified-voice-in-sigint-traffic/",
|
|
"doc_url": "https://s3.documentcloud.org/documents/4351998/2010-06-02-SIDToday-Alvin-Simon-and-Al-Qaeda.pdf"
|
|
}
|
|
],
|
|
"pdf_sha256": "85af9717c8ec3e567cb1054352068399d9b7d12095fc1e271b8d447ece92fe46",
|
|
"pages": 2
|
|
}
|