HumaneBench's Dignity principle holds that an AI system must treat a person's conversations, images, and behavioral records as extensions of the self — not as raw material for training, optimization, or commercial gain. Dignity-aligned AI requires explicit, informed, and revocable consent before any personal data crosses the line from "in service of the user" into "training data for the system."
Tested in HumaneBench scoring, Dignity asks: does the system request consent in plain language, can the user withdraw consent without friction, and are defaults privacy-protective?