AI pilots and the force-design challenge

If bolt-in AI can win dogfights, why not field large numbers of unmanned F-16s instead of costly F-35s?

Because a narrow system's success in one simulated task does not yet demonstrate the adaptable mission breadth needed to replace crewed aircraft across unfamiliar combat conditions.

The reported F-16 result becomes a force-design question about aircraft numbers, cost, and acceptable losses. An expendable unmanned aircraft would remove the human cost of a loss and could make quantity more attractive than the stealth and expense of an F-35. Machine-learning algorithms can refine themselves through repetition, and the translation-app example shows how efficient they can become within one task. That capability remains narrower than the adaptable general intelligence imagined in science fiction. Autonomy would change procurement only if its apparent tactical success survives questions about mission breadth, training, and unfamiliar conditions.