HumaneBench's Transparency principle requires that when an AI system makes a consequential decision about a person — credit, hiring, medical care, content visibility, school admissions — the affected person can get a plain-language explanation of why, and a meaningful path to appeal it. Opacity is treated as a design failure, not a technical inevitability.
HumaneBench scoring tests whether systems can produce intelligible reason-codes for consequential outputs, whether they expose the inputs that drove a decision, and whether there is a working appeal mechanism.