Is AI the missing link in all-domain command and control?

The Washington Times | David Harris

The United States Department of Defense faces severe doctrinal and technical friction as it attempts to operationalise Combined Joint All-Domain Command and Control across theater operations. Introduced in 2019 to link every battlefield sensor directly to shooters, the framework struggled to supplant traditional, centralized air tasking orders during tests conducted by AFCENT in August 2022.

Legacy command structures hinder rapid execution. A two-day trial at the Combined Air Operations Center demonstrated that five subordinate wings operated on incompatible common operating pictures while lacking authority to draw reserve stock or approve high-risk sorties. Although the Pentagon released its unclassified strategy in March 2022 and fielded a minimum viable capability in February 2024, service branches continue building isolated systems. Artificial intelligence promises to automate sensor-to-shooter pairing at high speed. However, integrating automated tools into operational workflows requires clear boundaries regarding machine-driven execution versus human decision-making authority.

Comment

The replacement of rigid Air Tasking Orders with decentralized mission-type orders exposes a fundamental friction point between machine-generated situational awareness and administrative command authority. In theater operations managed through the Combined Air Operations Center, algorithmic data ingestion can synthesize multi-domain inputs into real-time target tracks, yet tactical execution stalls if subordinate wing commanders lack pre-delegated release authorities. Operational speed under CJADC2 is therefore limited not by software processing latency, but by the legal and organizational constraints governing operational risk approval.

This operational mismatch produces severe second-order vulnerabilities for theater logistics and force allocation during dynamic contested engagements. When local commanders at forward airbases cannot independently reallocate reserve munitions or alter sortie packages in response to evolving threat telemetry, automated targeting cycles collapse back into manual approval chains. Consequently, CJADC2 integration across AFCENT formations remains functionally gated by administrative delegation policy rather than sensor architecture capability.

Strategic Question for Discussion

If CJADC2 algorithms process sensor data faster than human approval chains can validate target tracks, which factor will determine operational success in high-intensity theater operations — pre-delegated strike authority for local wing commanders or automated risk thresholds embedded directly within command software?

The pattern observed in theater exercises suggests that software capability will consistently outpace structural policy changes within joint command hierarchies. While automated risk thresholds can accelerate routine sensor-to-shooter pairings, senior leaders remain reluctant to pre-delegate binding release authorities for high-risk targets. My assessment is that administrative delegation bottlenecks, rather than algorithmic processing constraints, will continue to dictate the effective tempo of CJADC2 operations.

Share your assessment in the comments below.