The Controllability Trap: A Governance Framework for Military AI Agents

Researchers propose the Agentic Military AI Governance Framework (AMAGF) to address control failures in autonomous military AI systems. The framework introduces a Control Quality Score (CQS) to continuously measure and manage human control over AI agents throughout operations, moving beyond binary control models.

• →Six distinct governance failures are identified in agentic AI systems that current safety frameworks don't address.
• →The AMAGF framework uses three pillars: Preventive, Detective, and Corrective Governance to maintain human control.
• →A Control Quality Score (CQS) provides real-time metrics to quantify human control over AI systems.
• →The framework assigns responsibilities across five institutional actors with concrete evaluation metrics.
• →Military AI governance must shift from binary to continuous control models with active measurement.