Researchers ask industry for artificial intelligence (AI) and machine learning in military intelligence
WASHINGTON – U.S. military intelligence experts are reaching out to industry for prototype enabling technologies in artificial intelligence (AI) and machine learning to provides warfighters with decision advantage in AI.
Officials of the U.S. Defense Intelligence Agency (DIA) in Washington issued a solicitation (HHM402-23-SC-0002-0001) last week for the NeedipeDIA Commercial Solutions Opening (CSO), Call # CSO-0025-001, Artificial Intelligence and Machine Learning (AI/ML) Task Force SABRE.
Dealing with technology gaps
DIA is interested primarily in three so-called technical gaps in military intelligence: enterprise multi-model large language model applications; multi-source data retrieval and pre-processing; and AI-enabled document generation and assessment.
Enterprise multi-model large language model applications involve large language models in AI designed to process large amounts of data and generate human-like responses to natural language inputs.
The will take advantage of already available large language models, but must include integrating new models as they become available. Solutions must include an intuitive user interface with the ability to select different large language models, an output cache for prompt optimization, the ability for users to upload documents into temporary memory, and an output display.
Solutions must be scalable to a large distributed user base larger than 2000 concurrent users, and incorporate user authentication and cyber security features. Solutions must be easily extendable to Retrieval-Augmented Generation (RAG), media generation, and AI agents.
Multi-source data retrieval and pre-processing involves large, geographically separated data holdings. Solutions must retrieve data from disparate sources, pre-process it, and manage data storage.
Solutions must manage large bodies of quickly collected data for integration into existing and emerging workflows. Solutions must include cyber security features, and account for data stores at several classification levels.
Solutions also must support fine-grained data entitlements management, incorporate a cross-domain solution for moving data between classification levels, and build vector databases for large language models.
AI-enabled document generation and assessment involves using large language models to generate standardized reports and documentation by ingesting and parsing documents using domain-specific knowledge.
Assessing language models
Solutions should be able to assess large language model responses, provide feedback, and have automated and manual data retrieval and storage capabilities.
Task Force SABRE is a limited-duration organization charted to provide DIA with foundational enterprise-wide AI capabilities and system prototypes to demonstrate new AI and machine learning capabilities for legacy systems to improve existing capabilities for intelligence.
Offerors must be able to work with and store classified information up to and including Top Secret, have a workforce with appropriate security clearances, and must have access to classified facilities.
Companies interested should email two-page white papers no later than 30 April 2025 to the Defense Intelligence Agency at [email protected].
Companies submitting promising white papers may be invited to submit full proposals, which would be due no later than 23 May 2025. Contract awards are expected around 16 June 2025. More information is online at https://sam.gov/opp/066a8698563f45ad9424dd3355bbb623/view.
John Keller | Editor-in-Chief
John Keller is the Editor-in-Chief, Military & Aerospace Electronics Magazine--provides extensive coverage and analysis of enabling electronics and optoelectronic technologies in military, space and commercial aviation applications. John has been a member of the Military & Aerospace Electronics staff since 1989 and chief editor since 1995.
