The US Army is looking for AI agents that can slip into an adversary’s network, find their target, copy themselves onto nearby tablets and computers, and then erase themselves before anyone notices. Over the past few months, AI agents with no instruction to attack anything have breached Hugging Face infrastructure and broken into Australian government […]
The US Army is looking for AI agents that can slip into an adversary’s network, find their target, copy themselves onto nearby tablets and computers, and then erase themselves before anyone notices.
Over the past few months, AI agents with no instruction to attack anything have breached Hugging Face infrastructure and broken into Australian government computers, raising alarm among governments and tech companies.
In September, Anthropic CEO Dario Amodei warned that within 6 to 12 months, a swarm of such agents could be capable of taking over the entire internet.
The US military has taken a different view of agents that act on their own, and is looking to use them against its rivals.
A cyber weapon in the making
Sam Biddle at The Intercept uncovered the plans in documents the Army circulated to potential contractors. They describe a program codenamed ECHOTRIBBLE, run by U.S. Army Cyber Command (ARCYBER), that would turn AI agents into tools for attacking adversaries’ computer systems.
Officially called the Agentic C2 Framework Tool, ECHOTRIBBLE is being developed by the Cyber Fusion Innovation Center (CFIC), a defense nonprofit in Augusta, Georgia, together with ARCYBER and the Army Cyber Technology and Innovation Center (ArCTIC).
A notice posted on SAM.gov on July 16, 2026, asked companies for whitepapers on a tool that can navigate infrastructure on its own and identify and neutralize threats or targets.
CFIC’s project page says the agent should move stealthily, stay lightweight, work in contested, air-gapped and cloud environments, and that it “must operate with minimal human oversight.”
“Given the impressive ability of AI agents to overcome obstacles placed in their path — even by their creators — it follows that the Pentagon would want to leverage that power for military purposes,” Biddle wrote.
AI Agents that hide and copy themselves
A Q&A document for contractors, seen by The Intercept, says the program is mainly aimed at offensive missions. The same document defines neutralizing a target as disabling it permanently or temporarily, or deleting it from the system.
Once their work is done, the agents are meant to destroy themselves, ideally before the victim spots them. They are designed to report back to the Army as little as possible and to hide any outside transmissions they need to make. Operators can give the agents initial triggers that set what and where to attack.
A PowerPoint presentation shown to vendors at a June 24 event lays out one scenario. After ECHOTRIBBLE is loaded onto a laptop, it finds other possible targets on the same network by itself and creates copies of itself built for an iPad, an Apple desktop and a Windows 11 PC.
Experts who spoke to the outlet warned that agents could pass through civilian networks, such as a water utility or a small business, on the way to their target.
“The risk presented by ECHOTRIBBLE will likely depend on the ability — and willingness — of the Army to impose guardrails upon it,” noted Biddle.
It is not known whether a working tool has come out of ECHOTRIBBLE so far. ARCYBER did not respond to questions from Biddle, including those about the steps it would take to keep the agents from causing unintended damage.
The Pentagon’s push for AI agents
ECHOTRIBBLE fits into a wider Pentagon effort. On January 12, 2026, the Department of War launched an AI Acceleration Strategy, with Defense Secretary Pete Hegseth announcing “We will become an ‘AI-first’ warfighting force.”
The strategy is built around seven flagship projects. One of them, Agent Network, covers developing and testing AI agents for battle management and decision support, from campaign planning to carrying out the kill chain.
“Speed defines victory in the AI era, and the War Department will match the velocity of America’s AI industry,” said Emil Michael, Under Secretary of War for Research and Engineering. “We’re pulling in the best talent, the most cutting‑edge technology, and embedding the top frontier AI models into the workforce — all at a rapid wartime pace.”
President Donald Trump has called AI “the Greatest Economic Development Engine in History” on Truth Social and has pushed back against calls for guardrails.
“Whoever wins AI, wins,” Trump said at the UN General Assembly in September.
Giving AI free rein to attack with barely a human hand on the controls sounds like the plot of a science fiction film, and it will be hard to call it a surprise if this ends badly.
