WASHINGTON -- The Army continues to refine and evaluate an artificial intelligence software prototype for Army intelligence that is designed to quickly identify threats through a range of battlefield data and satellite imagery.
"Prometheus" is a machine learning capability designed to recognize potential threats and provide targeting coordinates through the rapid analysis of intelligence information, said William Nelson, director of the Assured Position, Navigation and Timing Cross-Functional Team.
The prototype software will integrate with the Army's Tactical Intelligence Targeting Access Node, or TITAN, Nelson said during this year's Association of the U.S. Army Annual Meeting and Exposition on Oct. 15.
"Our vision for the space domain is to ensure the Army can leverage the entirety of space-based [intelligence, surveillance and reconnaissance]," said Lt. Gen. Laura Potter, the Army deputy chief of staff, G-2.
Providing commanders with a tactical or strategic ISR advantage will allow them to see, understand, decide, and act faster, thereby reducing the decision-making process from minutes to seconds, Potter added.
"TITAN leverages data, products and services" from the land, air, sea, space, and cyber domains, Potter said. It provides the Army with a "scalable expeditionary and tactical ground station … that integrates Department of Defense, intelligence community, and commercial data, products and services."
The APNT CFT, which falls under Army Futures Command, along with the Intelligence, Surveillance, and Reconnaissance Task Force recently evaluated Prometheus and a TITAN surrogate during Project Convergence 20, team members said. The demonstration considered the Army's ability to sense and identify targets at extended ranges through the employment of space-based capabilities and AI data analytics.
Nelson said both systems were capable of passing target information to an Extended Range Artillery Cannon during the evaluation. In the future, program officials plan to extend Prometheus' reach, furthering its ability to collect data from various ground and air systems.
The Army is also working to implement anti-spoofing measures into its AI software to ensure targets are properly identified, said Alexander Miller, science and technology senior advisor with the Army’s G-2 office.
“If I am somehow seeing a tank on the water … I know that is probably wrong,” Miller said. “I have to make the system smart enough to fact check [the data].”
To prevent spoofing, officials must ensure the security and fidelity of root data, all while scrutinizing AI outputs to develop trust, Miller said. He added that program officials will need to take the necessary steps to analyze how AI incorporates incoming data and utilizes it during the process.
Continued efforts by the Army's ISR and AI task forces and APNT officials will enable the force's ability to detect, characterize, and take action at the speed of demand, Potter said. Doing so will meet the Army's multi-domain operational requirements during times of competition, crisis, or conflict.
"Our strong and enduring partnerships across the Department of Defense and the intelligence community ensure that Army intelligence leverages the power of the enterprise to support the Army's modernization strategy," Potter added.