Army Seeks Digital Twin of Its Network to Boost Training, Testing, and Cyber Defense, NETCOM Chief Says

WASHINGTON — To fully harness artificial intelligence, the Army requires a comprehensive simulation of its networks — a digital twin — where both algorithms and the personnel operating alongside them can be trained and tested, the head of the service’s Network Command (NETCOM) stated today.

Maj. Gen. Jacqueline Denise McPhail acknowledged at AFCEA’s annual TechNet Augusta conference that building such a detailed, continuously updated digital model presents significant difficulty. However, she emphasized its vast potential, challenging industry contractors in the audience to undertake the effort regardless.

The primary benefit, McPhail explained, would be an ultra-realistic training environment for network operators, cyber defenders, and AI systems to understand network operations and maintain functionality even while under attack.

“We get about 1.2 million cyber attacks a day,” the NETCOM commander said, noting that these threats are increasingly sophisticated and far harder to detect than traditional brute-force Distributed Denial of Service (DDoS) attacks.

“It’s no longer DDoS. Now we have to look at behavior: What behavior has changed?” she said. “We face a tremendous volume of incoming data, and the challenge is how we secure that data and distinguish between mere noise and an actual shift in behavior.”

“That’s where AI comes in,” McPhail continued. “If we can develop a digital twin of the DoDIN-A — the Department of Defense Information Network–Army, encompassing all military networks for which the Army is responsible — we can leverage AI within that model to identify vulnerabilities, pinpoint gaps, and locate areas of resilience.”

“It’s a big ask. It’s a big ask! I know it’s a large-scale effort,” she conceded. “I think we start small and build it out.”

McPhail is applying a broader definition of “digital twin” than the one codified in Pentagon doctrine, which defines it as a real-time digital counterpart of a physical object or process. Typically, defense officials and industry leaders use the term to describe highly detailed simulations of physical assets — aircraft, supply chains, or virtual patients for medic training — replicating components and interactions so accurately that the virtual model mirrors the physical system’s behavior in any scenario. This allows for stress-testing, design iteration, and predictive analysis without risking the actual asset.

McPhail argues that the same advantages can be achieved for a system that is already largely digital: the Army network. While the network relies on physical hardware, the critical activity worth modeling occurs at the software layer.

“We need to figure out what’s broken, but we also need to figure out what’s not broken so we don’t break it later,” she added.

Source link