When most people imagine artificial intelligence transforming modern warfare, they picture gleaming data centers, satellite feeds, or high-altitude drones. According to Military Embedded Systems, that picture misses where the most consequential change is actually unfolding: at ground level, in the dirt.
The publication argues that the true AI revolution in defense is advancing through autonomous ground systems — robots, unmanned vehicles, and embedded processors operating in close contact with terrain, troops, and the enemy. This is a quieter story than the headlines about AI-assisted targeting or Pentagon cloud contracts, but the argument is that it may be the more operationally significant one.
Ground-level autonomous systems face challenges that airborne or networked platforms do not: contested communications, unpredictable physical environments, and the need for rapid, low-latency decisions without a human in the loop for every action. Solving those problems, according to Military Embedded Systems, is where defense-grade AI is being stress-tested and refined in ways that matter most for battlefield outcomes.
The framing — "happening in the dirt" — is a deliberate contrast to the perception that military AI is primarily a software or cloud story driven by large tech contractors. The real engineering, the argument goes, is embedded, ruggedized, and rolling across uneven ground.
Why it matters: as great-power competition intensifies, the nations that figure out reliable autonomous ground systems soonest may hold a decisive edge in the kind of attritional, terrain-contested warfare that recent conflicts suggest is far from obsolete.