AI Over-Trust and Diminishing the Joint Targeting Cycle: Why Doctrinal Discipline Matters More Than Artificial Intelligence

Small Wars Journal | Jesse R. Crifasi

Project Maven and other artificial intelligence-enabled decision-support systems are rapidly accelerating the Joint Targeting Cycle, but they risk eroding critical human judgment across the modern battlefield. This technological integration compresses planning timelines, tempting commanders to bypass established doctrinal processes for faster engagements, which ultimately compromises meaningful human control.

Historically, the six-phase targeting methodology relied on deliberate human analysis to translate strategic objectives into precise military actions. However, algorithmic mediation now pre-filters intelligence, introducing compounded cognitive uncertainty before recommendations reach commanders. Automation bias further encourages staffs to accept these machine-generated outputs without sufficient scrutiny. Speed does not improve reasoning. To preserve accountability under DoD Directive 3000.09, the Joint Force requires strict doctrinal discipline, ensuring human discernment actively interrogates algorithmic assumptions throughout target development rather than merely at the final point of weapon release, thereby preventing flawed machine logic from dictating lethal operational outcomes on the battlefield.

Comment

The integration of AI-enabled decision-support systems into the Joint Targeting Cycle challenges the core tenets of DoD Directive 3000.09. By algorithmically mediating intelligence before it reaches human operators, these systems subtly pre-determine target validation and prioritisation. This shift moves critical military judgement from the analytical phase to the final point of engagement. Consequently, the deliberate, phase-by-phase evaluation of assumptions mandated by joint doctrine is replaced by a compressed process that prioritises speed over analytical depth.

This doctrinal erosion is particularly evident in the deployment of Project Maven, where computer vision algorithms automate the initial detection and classification of potential threats. When these automated outputs feed directly into subsequent target-pairing databases, the underlying algorithmic assumptions are rarely re-evaluated by human analysts. The resulting targeting recommendations carry an artificial veneer of certainty that undermines the rigorous verification standards established under the Joint Targeting Cycle during Project Maven operations.

Strategic Question for Discussion

If Project Maven continues to automate target detection and classification at the tactical edge, how can commanders ensure that the appropriate human judgment mandated by DoD Directive 3000.09 remains substantively active during the initial target development phases?

The available evidence points toward a structural tension between rapid machine processing and the deliberate verification required by joint doctrine. Resolving this tension likely depends on integrating explicit human validation milestones within the target-pairing database architecture. This trajectory indicates that maintaining substantive compliance with DoD Directive 3000.09 will rely on engineering deliberate analytical friction back into the automated workflow.

Share your assessment in the comments below.

To read in details click here