snowchive-meta/data/black-budget-262/metadata.json
2026-11-19 04:39:11 +02:00

58 lines
4 KiB
JSON
Raw Permalink Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

{
"title": "“Black Budget” — FY 2013 Congressional Budget Justification/National Intelligence Program, p. 262 (February 2012)",
"type": "REPORTED",
"description": "This excerpt from the Fiscal Year 2013 Congressional Budget Justification (the \"Black Budget\") details the NSA's financial and performance metrics for \"Analytics Modernization,\" specifically focusing on the transition to high-volume, automated speech-to-text systems. It records the decommissioning of legacy voice processing systems (VoiceRT) in favor of more robust, scalable capabilities aimed at rendering millions of intercepted voice files searchable via text queries and paired dialogue transcription.",
"long_description": "The document provides a rare financial window into the industrialization of the NSA’s \"bulk listening\" capabilities. A primary metric, labeled CCP_00083, tracks the number of voice files rendered searchable by text queries. While the agency missed its FY 2011 target of 350,000 files—finishing at approximately 285,337—the failure is attributed to a strategic transition. The NSA decommissioned VoiceRT systems outside of Washington to prepare for SPIRITFIRE, a more advanced voice analytic system operationalized by 2013 that focuses on automated keyword searching and full dialogue transcription.\n\nFinancially, the \"Analytics Modernization\" project received significant backing, with funding peaking at $512.5 million in FY 2012. A key efficiency metric, CCP_00103, shows the agency’s success in drastically lowering the \"average cost to deliver an analytic hour equivalent.\" By leveraging machine translation (MT) and automated speech recognition (ASR), the NSA reduced the cost per analytic hour from $1.12 in 2008 to a projected $0.21 by 2013. This shift signifies a move away from expensive human linguistics toward a high-speed, machine-driven SIGINT enterprise.\n\nThe budget also highlights infrastructure and human capital challenges. Project CCP_00169 details the struggle to establish a massive \"enterprise distributed database architecture\" for content storage, which faced delays due to engineering changes and facility leasing issues. Additionally, metrics for the Linguists/Translators project (CCP_00179) show that while the number of analysts trained in Digital Network Intelligence (DNI) increased, high attrition rates and an intense \"operations tempo\" hindered the agency from reaching its full training benchmarks. This emphasizes that even as the NSA pivots toward automation, the demand for specialized human analysts to oversee these \"Black Budget\" systems remains a bottleneck.",
"tags": [
"Black Budget",
"analytic modernization",
"speech-to-text",
"speech recognition",
"machine translation",
"congressional budget",
"voice intercept",
"keyword search",
"dialogue transcription",
"Human Language Technology",
"automated surveillance",
"linguistics",
"language"
],
"document": {
"creator": "NSA",
"date_created": "2012-02-01",
"classification": [
"TOP SECRET",
"SI",
"TK"
],
"dissemination": [
"NOFORN"
],
"banner": "TOP SECRET//SI/TK//NOFORN",
"genre": "other",
"programs": [
"VoiceRT"
],
"countries": [],
"people": []
},
"publication": [
{
"publisher": "The Intercept",
"date_published": "2015-05-05",
"title": "How the NSA Converts Spoken Words Into Searchable Text",
"link": "https://theintercept.com/2015/05/05/nsa-speech-recognition-snowden-searchable-text/",
"reporters": [
"Froomkin, Dan"
],
"contributors": [
"Fishman, Andrew"
],
"doc_source": "https://theintercept.com/document/2015/05/05/black-budget-fy-2013-congressional-budget-justificationnational-intelligence-program-p-262/",
"doc_url": "https://s3.documentcloud.org/documents/2072013/black-budget-262.pdf"
}
],
"pdf_sha256": "73701e5015a3c7e27951ba5f9d3748a756c25143b8c7aaa3e62c5037ba1b24b5",
"pages": 1
}