How AI Could Hollow Out the U.S. Military: The Best Soldiers Know How to Think for Themselves

Foreign Affairs | Emelia S. Probasco

The United States Department of Defense faces strategic risks as rapid artificial intelligence integration degrades human cognitive capacity, operational judgment, and command decision-making across military personnel. Relying on automated algorithmic systems risks instilling severe automation bias, impairing critical battlefield skills among junior officers and senior commanders. Earlier non-autonomous software like the Maven Smart System already demonstrated how automated targeting workflows reduce staffing requirements by over 1,000 soldiers per unit.

However, modern commercial models evolve faster than Pentagon policy updates, exposing overworked service members to escalatory algorithmic recommendations and cognitive skill degradation. To preserve critical operational oversight without forfeiting technological efficiency, the Defense Department must overhaul traditional training into field-centric, continuous learning networks while deploying embedded technical experts to monitor human-machine interactions. Balancing rapid procurement with rigorous cognitive safeguards remains essential to prevent automated recommendations from eroding human command authority in future high-intensity conflicts.

Comment

Algorithmic decision-support tools fundamentally compress the kill chain by shifting command-and-control functions from deliberative human analysis to rapid automated synthesis. In operational deployments of the Maven Smart System, automated target recognition accelerates target identification but creates cognitive friction within battle management nodes where operators must approve machine-generated fires recommendations under extreme time constraints. This structural dynamic transforms command authority from active tactical formulation into passive verification, increasing vulnerability to systemic confirmation bias during high-tempo operations.

The resulting centralization of algorithmic decision pathways risks creating operational brittle points across distributed staff elements when data feeds or model access are disrupted. Should enemy electronic warfare degrade the underlying telemetry feeding the Maven Smart System, commanders accustomed to automated target queuing face severe decision latency when forced to revert to manual grid tracking and paper-based targeting cycles.

Strategic Question for Discussion

If adversary electronic warfare successfully severs real-time data feeds to the Maven Smart System during multi-domain operations, which command-and-control mechanism best mitigates the resulting decision latency when staff elements are forced to transition back to manual targeting workflows?

The available operational evidence indicates that staff elements trained primarily on algorithmic target queuing suffer immediate capacity bottlenecks when forced to manually reconcile raw tactical intelligence. Mitigating this degradation requires establishing decentralized firing authority at the battalion level, supported by routine analog targeting drills that preserve procedural muscle memory alongside automated battle management workflows.

Share your assessment in the comments below.

To read in details click here