The U.S. military's cyber warfare division is preparing to field artificial intelligence systems with one directive: find the most powerful models available, regardless of where they come from or who built them.
Brig. Gen. Reid Novotny, who leads AI strategy for U.S. Cyber Command, told reporters the military branch is building technical infrastructure designed to swap between competing AI systems as battlefield conditions demand. That flexibility includes the possibility of deploying open-source models or systems developed in adversarial nations, if they prove militarily superior.
"To survive anywhere, just in case our operators want an open-source made-in-China model or something very boutique, we have to create the infrastructure and that ability to be agile," Novotny said. "There are no politics in this."
The stance puts Cyber Command at odds with broader Pentagon tensions over AI vendor selection. Anthropic, whose models are considered among the most capable, has been locked in a dispute with the Defense Department over access to its latest systems. The company has restricted government deployment of its Mythos Preview over concerns about its offensive hacking potential. Only scattered federal agencies, including the National Security Agency, currently have clearance to use it.
Meanwhile, OpenAI is aggressively courting government contracts with its GPT-5.4-Cyber system, while Google has already secured agreements to integrate its Gemini models into classified Pentagon operations.
Cyber Command's approach sidesteps this vendor competition by building toward what Novotny describes as a vendor-agnostic operating system for military AI deployment. For the first time in 2026, the command has dedicated funding specifically for AI programs after years of pushing through Pentagon and congressional appropriations processes.
That money is being used to pilot commercial AI capabilities while establishing underlying systems that allow operators to migrate between different models as technology evolves. "Our operators are very well set for what they need right now," Novotny said.
Novotny, the first AI officer appointed to Cyber Command, is tasked with integrating artificial intelligence into both offensive and defensive cyber operations. The goal is to accelerate attack capabilities and process massive volumes of intelligence data while managing the inherent risks of autonomous systems in combat zones.
The most contentious concern is civilian harm. Cyber attacks on critical infrastructure like hospitals or power grids could kill people. Novotny argued the military already has rules in place to govern those risks and sees no need for AI-specific restrictions. When training a system to breach a nation's networks, he said, "we know going in about hospitals, schools and so on." The military validates that models behave as intended before deployment.
"I don't want to say we're not worried about it, but we know how to apply our morals and our laws when we adopt a new technology," Novotny said during a panel discussion.
Cyber Command is experimenting with different levels of human oversight in AI operations, but fully autonomous deployment without any human in the decision loop is off limits. "We would never unleash a human-out-of-the-loop tool and then be like, 'Oops, we just turned something on,'" Novotny said.
The bigger practical challenge, according to former military cyber leaders, is speed. Lt. Gen. Lori Reynolds, who previously commanded Marine Corps cyber forces, warned that AI capabilities are likely to overwhelm the Pentagon's existing operational model, which was built around engagements with a small number of sophisticated adversaries.
"It's insufficient by itself," Reynolds said. "We have to learn to fight hurt, we have to learn to trust our data, but we should be doing everything we can to constantly be sparring as far away from our territory as we can."
The military's AI push is unfolding as rival powers race to gain advantage in cyber warfare. China is integrating AI into its own operations. Competitors in the private sector are racing to secure Pentagon contracts. The stakes for speed and capability are climbing fast.
Author James Rodriguez: "Novotny's 'no politics' message sounds noble until you realize the Pentagon is already picking winners and losers among AI vendors through classified contracts, just less transparently."
Comments