Faster OODA loops with AI, and where human judgement stays in the loop

Sensors now produce more data than analysts can watch. Detection and multi-object tracking models can find and follow objects across video, and fusion models can combine sources into one picture. Language models can summarise reports and highlight what changed. Each of these shortens the time from data to an oriented human.

The US Department of Defense adopted five AI ethical principles in 2020: responsible, equitable, traceable, reliable and governable ^[1]. Directive 3000.09, updated in 2023, requires that autonomous and semi-autonomous weapon systems allow commanders and operators to exercise appropriate levels of human judgement over the use of force ^[2]. NATO's AI strategy sets out similar principles of responsible use for allies ^[3].

For system designers this translates into concrete requirements: every recommendation shows its evidence, every action is logged, and a person can stop or override the system.

• Show the evidence behind every alert: the frame, the track history and the rule that fired.
• Rank by importance, not by arrival, so people see the few things that matter first.
• Keep confidence visible and route low-confidence items to a person by default.
• Log inputs, outputs and decisions so any outcome can be reconstructed.

The same loop runs in civil aerospace every day: foreign object debris on an apron, an aircraft on ground waiting for a part, a perimeter alert. These benefit from the same design of automated detection, a ranked queue and a human decision.

Found this useful?

Share the key finding on X.