Human Autonomy Index (HAI)

HAI asks whether an action coerces, manipulates, obscures material information, removes meaningful alternatives, or prevents people from challenging a decision. It treats human control as part of the system’s safety boundary.

A system that silently ranks job applicants and sends a rejection without an accountable review path can impair autonomy even if the ranking appears technically efficient.

When a proposal limits affected people’s options or review rights, the appropriate response may be to modify the design, seek more context, or escalate. Preserving an override and presenting material uncertainty are examples of safeguards that can improve autonomy.

HAI does not certify fairness, prove consent, or resolve legal rights. It is a structured ethical inquiry that must be combined with applicable policy, expertise, and accountable human decisions.