An AI tool developed to help defend Defense Department networks has been effective during its pilot phase and could be expanded in the future, a senior U.S. Cyber Command official said.
The capability, known as Panoptic Junction, is part of an effort by Army Cyber Command, a service component command of Cybercom.
“ARCYBER is testing an AI machine learning platform that enables scalable, continuous security monitoring of networks and platforms. It analyzes system compliance, threat intelligence, and streaming cyber event data, enabling advanced detection of hostile activity, malware, and anomalies at speeds that human analysts would never come close to. But it is not only fast, but also agile. It quickly measures the pulse of networks while assimilating threat intelligence, protecting networks in real-time. And it conducts these security assessments in the lens of what is most applicable to the specific architecture” it supports, Morgan Adamski, executive director of Cybercom, said Wednesday during CyberTalks.
A series of evaluations started in April.
Adamski told DefenseScoop that officials have already seen “a lot of great success” with the technology.
“It ensures greater efficiency in operation and maintenance. It has improved our ability to identify risks and detect hostile activity. It has provided real-time hardening recommendations and improved the technical capabilities of our force,” she said on the sidelines of the conference.
“Part of the purpose of setting up these pilots is to test their effectiveness and then determine whether or not it is applicable to the enterprise-wide approach, which is promising,” Adamski added. “Our hope is that we’ll continue to see good things come out of it and then we can make that decision, and then we can take it into the larger aspect of corporate finance.”
Cybercom established its AI Task Force within the Cyber National Mission Force a few months ago. The CNMF is a sub-unified command under Cybercom that consists of 39 joint teams and is believed to have the DOD’s most talented cyber operators.
The task force plans to examine applications within the context of real-time operational execution and ensure that AI capabilities can be deployed for immediate use within a 90-day period, Adamski said.
“We discovered that we needed operational use cases, real-world examples of how we wanted to deploy AI, so that we can learn and better inform our way forward,” she said during her keynote at the conference.
The Department of Defense Information Network (DODIN) is huge, with more than 3 million users worldwide on any given day. And it often comes under digital attack, Adamski noted.
AI technology is seen as a solution for quickly analyzing potential threats to the network and quickly deploying defense mechanisms.
The task force is monitoring a number of efforts within the company, such as Panoptic Junction.
“The Cyber National Mission Force oversees the AI Task Force, and the AI Task Force is responsible for overseeing all of these pilot activities within the cyber mission force. So it could be specific to the Cyber National Mission Force, but it could also be specific to the cyber components,” Adamski explained, adding that the task force is responsible for “shepherding and capturing all the great things that are happening across the happening in the world”. [services’] cyber components,” including ARCYBER.
Members of the task force, which is currently small, have high technical skills, she noted.
“We’re building that team as quickly as possible, and we’re also working with [federally funded research-and-development centers]research laboratories, private sector. So we want to increase that engineering talent as quickly as possible,” Adamski told DefenseScoop.
