Army wants AI to improve analysis of intelligence data
Connecting state and local government leaders
The Army wants to use artificial intelligence and machine learning to decrease the cognitive burden on intelligence analysts, provide timely information to decision-makers and identify specific targets with a high confidence in real time or near real time.
With all the complex, high-speed data coming at military intelligence analysts, decision-makers and warfighters, the Army wants to use artificial intelligence and machine learning to decrease the cognitive burden on analysts, provide timely information to decision-makers and identify specific targets with a high confidence in real or near-real time.
As part of its intelligence modernization efforts, the Army plans to provide three interdependent, AI-enabled systems that improve processing and data dissemination capabilities for intelligence analysts across multiple intelligence domains on various security classification networks.
- A new data warehousing capability will include data ingress, cleansing, processing, egress and querying to enable storage and timely processing, exploitation and dissemination of intelligence data.
- A ground station modernization program will leverage modern networking technologies for data discovery, data exchange and the use of AI/ML applications that improve the timeliness and accuracy of reports to combat operations even in disconnected environments. The near-term focus is on incorporating data from national, joint, commercial and Army sensors and platforms.
- A new software baseline called Intel Apps will replace select legacy capabilities in the command post computing environment, allowing analysts to process intelligence faster and more accurately.
In an April 27 request for information, the Army wants to hear from providers who can deliver explainable AI/ML models for processing a variety of applications, including moving target indicators, electronic intelligence, measurement and signature intelligence and natural language processing. It also wants to know what kind of sensors are supported, whether the tools are open source or proprietary and the type, amount and veracity of training data required for the models. Responders are also asked to provide security and company/cost details.
Responses are due May 14.