Offline Learning and Counter-AI for Autonomous Aircraft Combat Operations (AACO/TACFI) — a federal contract (USAspending)

This contract centers on operating aircraft that act autonomously, without a human pilot, in combat settings (AACO, Autonomous Aircraft Combat Operations), and it names two technical themes. The first is "offline learning," which means completing the learning ahead of time using data prepared in advance, rather than learning on the spot in the field. In combat environments where communication may be constrained or split-second decisions are required, there is little room to learn slowly in the moment, so the underlying idea is to bring in a decision model that has already been trained. The second is "counter-AI," which assumes the other side also uses AI and refers to techniques for deceiving or neutralizing an adversary's AI, or otherwise countering the decisions an adversary's AI makes.

The contract matters because it directly addresses a shift in how AI is used, from "human versus machine" toward "machine versus machine." Once one side automates its judgment with AI, the other side begins to treat that automated judgment itself as a target to overcome. In other words, the question is no longer only how to make an AI smarter, but how to design an "AI that operates against other AI." Tackling both a preparation-through-prior-learning approach and an approach for countering and disrupting an adversary's AI within the same effort fits an outlook shaped by this kind of competitive environment.

Because the awarding party is the U.S. Air Force and the work proceeds under TACFI (Tactical Funding Increase), an additional-funding mechanism, the contract reads as one that adds money to an existing line of work to accelerate development. What technical foundation TOYON RESEARCH CORPORATION brings, and what was ultimately produced under this contract, are not stated in the original record. Even so, it is worth referencing across topics as a clue to where public procurement around autonomous systems and AI safety and countermeasures is directing its funding.