CORTEX: GAMBIT DEFENSE builds an AI/ML "opposing force" for counter-drone training for the U.S. Air Force — a federal contract (USAspending)

Drones (small unmanned aircraft) are inexpensive and widely available, and they have grown into a more prominent threat in military and security settings. Efforts to address them are known as "counter-unmanned aerial systems" (counter-UAS), a general term for the technologies and operations used to detect, identify, and neutralize hostile drones. Responding accurately to drone threats in the field depends on units training repeatedly. The role of the adversary in such training is filled by an "opposing force" (a unit that acts as the simulated enemy in an exercise), and this contract seeks to build that adversary function using artificial intelligence and machine learning.

As for why AI/ML is used, it is generally noted that when a training adversary behaves predictably and monotonously, it tends to fall short as preparation for the varied threats of the real world. By embedding AI/machine learning (systems that learn patterns from data and adapt their behavior to the situation) into the opposing force, a more varied and realistic adversary can be expected within the training environment. The project name "CORTEX" is the identifier for this effort; this direction can be read from the original text, but what performance or outcomes were actually achieved is not stated in the original.