The U.S. Army has built its first AI chatbot for soldiers, training it on classified data from active combat operations in Ukraine and Iran in a move that signals a decisive break from the Pentagon's reliance on commercial AI tools.
In an exclusive WIRED interview with Alex Miller, the Army's chief technology officer, unveiled a prototype of the system, called Victor. The platform pairs a Reddit-style peer forum with an AI chatbot called VictorBot, designed to help troops retrieve mission-specific knowledge without having to file formal requests through the military chain of command.
Miller told the publication that the data pool feeding the system already spans more than 500 repositories, drawing from active theatres including the Ukraine-Russia war and Operation Epic Fury, the U.S. air campaign against Iran.
How VictorBot Works: Peer Knowledge, Cited Sources, Machine Speed
Victor's architecture is straightforward in concept but novel in military application. Soldiers post questions and observations to a forum that functions like Reddit, where other service members can reply with guidance drawn from direct field experience. VictorBot, the AI layer, reads across those posts, synthesises an answer, and cites the specific threads it drew from, so the soldier can trace the source. 'Electromagnetic warfare is such a hard topic,' Miller told WIRED. Victor, he said, 'can generate a response and cite all of the lessons learned from [different] units.'
That sourcing mechanism is deliberate. Miller said the system is designed to reduce the risk of AI errors by grounding every response in verifiable Army records, an approach that mirrors the citation practices of commercial large language models but applies them exclusively to classified military data. The practical case Miller demonstrated was configuring electromagnetic warfare systems for a specific mission, a task that would previously have required a soldier to track down the right specialist or dig through disconnected field manuals.
Lt Col Jon Nielsen, who oversees Victor's development within the Combined Arms Command, told WIRED that brigades routinely repeat the same tactical mistakes across different deployments because knowledge stays siloed within units. Victor's goal, he said, is to break that cycle.
Nielsen also confirmed the system's planned expansion into multimodal inputs: 'Victor will be one of the only sources with access to authoritative Army information,' he said, adding that soldiers will eventually be able to upload imagery and video for the system to analyse.
Live Combat Data as Training Material: Ukraine, Iran, and Operation Epic Fury
What separates Victor from its commercial counterparts is the data feeding it. Miller confirmed to WIRED that the system is trained on operational intelligence gathered from the Ukraine-Russia war and Operation Epic Fury, the U.S. military designation for its air campaign against Iran.
That kind of training material, drawn from active conflict rather than open-source text, gives VictorBot a specificity that no commercially available model can match. The same classified environment that makes the data valuable also makes it impossible to feed into third-party platforms without risking a national security breach.
The Pentagon has previously relied on commercial AI tools, including Anthropic's Claude, for some planning and analysis functions. According to WIRED, Anthropic's technology played a role in planning operations in Iran through a Palantir-built system, though that arrangement became the source of a significant public dispute.
Anthropic argued publicly that its models should not power autonomous weapons or be used to surveil American citizens, a position that put it in direct conflict with the Department of Defence. Victor represents the Army's answer to that dependency: build and control the tool entirely in-house.
Miller told WIRED that the Army is working with an unnamed third-party vendor to run and fine-tune the underlying AI models, but declined to name the firm because the contract had not yet been publicly announced. The late last year launch of GenAI.mil, the Department of Defence's initiative to accelerate AI adoption across its workforce, provides the institutional backdrop for Victor. The Army has moved faster than most branches in converting that mandate into an operational prototype.
Expert Caution: Sycophancy, Agentic Risk, and the Limits of Military AI
Not everyone views Victor's development without reservation. Paul Scharre, executive vice president of the Center for a New American Security and a former U.S. Army Ranger, told WIRED that the tendency of AI models to be sycophantic poses a particular hazard in military contexts. 'I could envision situations where that would be particularly worrisome in a context of intelligence analysis,' he said.
Sycophantic AI, a well-documented behaviour in large language models, refers to the tendency of these systems to tell users what they want to hear rather than what is accurate, reinforcing existing assumptions rather than challenging them.
That risk compounds as AI systems grow more capable. Scharre noted that the transition from chatbots to agentic AI, systems that can independently take actions across software and computer networks, introduces an entirely new security surface. 'Agentic AI raises this whole new set of challenges around security,' he told WIRED. A February 2026 paper from Princeton found that default chatbot interactions resemble the cognitive effects of confirmatory evidence, increasing user confidence without bringing them closer to truth, a dynamic that carries obvious implications for targeting decisions and battlefield assessments.
Lauren Kahn, a senior research analyst at Georgetown's Center for Security and Emerging Technology and a former policy adviser in the Office of the Under Secretary of Defence for Policy, takes a more measured view. She told WIRED that Victor's clearest near-term value lies in automating administrative and logistical work within the DoD rather than frontline decision-making.
'The big labs are obviously going to have a comparative advantage,' she said, suggesting that if Victor matures, the Army may eventually bring in a major AI company to advance it further. Kahn also confirmed to WIRED that the project highlights how AI could free up significant human capacity from back-office functions that currently consume military personnel time.
VictorBot is still a prototype, but the Army's decision to build it on classified combat data rather than commercial infrastructure marks the clearest signal yet that AI is becoming as fundamental to military readiness as the weapons systems it is designed to support.
Originally published on IBTimes UK
© Copyright IBTimes 2025. All rights reserved.
