The U.S. Army Cyber Command is training artificial intelligence agents to perform specialized cyber roles alongside human personnel, with the goal of conducting network operations at machine speed.
Lt. Gen. Christopher Eubank, commander of Army Cyber Command, said Tuesday at the TechNet Augusta conference that every AI agent is trained for a specific role within the cyber force.
“We’ve created developers, data engineers, host analysts, exploitation analysts. You name the work role, we’re creating those,” Eubank said.
According to Eubank, the agents are trained to standards comparable to those applied to human personnel. They receive job qualifications, are evaluated for readiness and are assigned to missions under human supervision.
When an agent makes a mistake, operators correct it and teach it the proper procedure before sending it back for additional training.
Task Force Lexington Leads Army AI Efforts
The work is being coordinated through Task Force Lexington, which serves as a central hub for Army Cyber Command’s artificial intelligence initiatives.
The task force is led by an officer expected to be promoted to colonel and includes roughly 10 specialists whom Eubank described as “high-speed folks.”
AI agents are already performing critical tasks, including hunting across networks for cyber threats and intrusions. Other AI tools are being used for cybersecurity service-provider support, Risk Management Framework processes and the Enterprise Mission Assurance Support Service.
Army Cyber Command also has an agentic red team and 17 cyber protection team mission elements focused on monitoring and defending the network each day.
Army Seeks to Match Adversaries’ Machine Speed
Lt. Gen. Jeth Rey, the Army’s deputy chief of staff for communications, G-6, said the United States must accelerate its cyber operations to counter adversaries that are already using automation to operate faster than human teams.
“Defending and operating the network is definitely a thing that humans are doing right now. We’re doing it at human speed,” Rey said during the same panel. “But our adversaries are doing it at machine speed, and we must catch up and we must get ahead.”
Rey described the Army’s network as a critical warfighting system on which nearly every other military function depends.
“We can’t fire rounds without it, we can’t move people without it, we can’t operate on a daily basis without it,” he said.
Humans Continue to Control Risk Decisions
Despite the expanding use of AI agents, Eubank emphasized that the systems are not authorized to independently assume operational risk.
Each day, Army personnel establish the guardrails governing the agents and determine which risks must be handled by humans and which can be evaluated by artificial intelligence.
“We have not turned any agents loose to assume risk on their own behalf,” Eubank said. “But this is a monumental point as we operate, secure, defend and attack the network.”
Eubank described the process as a “delicate dance” because the agents move much faster than humans. The systems check in with human counterparts, who review their work and direct them back to specific tasks when further analysis or correction is required.
Engineers Receive Additional Training
To strengthen the team’s technical expertise, Eubank recently sent the entire Task Force Lexington staff to the Defense Innovation Unit, where its members met with senior engineers.
The initiative reflects the Army’s effort to develop not only AI agents, but also the specialists needed to design, train, supervise and integrate autonomous systems into military cyber operations.
