Definition
The rate of agreement between AI-generated battle damage assessments and independent human-only assessments, compared to chance, used to detect confirmation bias.
Formula
BDAII = Cohen kappa (AI-BDA vs independent human BDA); Bias_direction = mean(AI_score - Human_score)
Proposed formula, pending validation pilots.
At a Glance
Category
AI Calibration & Trust Integrity
Decision Supported
Can we trust the AI's target destroyed assessment when the same AI nominated the target?