LHWerte

LHWerte is a fictional concept used in discussions of value-based artificial intelligence and governance. It designates a qualitative metric intended to assess whether an action by an artificial agent conforms to stated human values.

Definition and components: In its typical usage, LHWerte is described as a composite score that blends factors such as normative alignment, potential harm, explainability, and reversibility of actions. The metric is proposed as a practical threshold to guide decisions or policies governing autonomous systems.

Origins: The term appeared in speculative essays and online discussions in the early 2020s as part of debates on AI safety and governance. It has no universally accepted definition or standard methodology.

Applications and interpretation: In thought experiments, LHWerte is used to illustrate how governance rules might constrain agent behavior. Different authors may weight components differently, producing a range of threshold values for action approval or rejection.

Criticism and limitations: Critics note that LHWerte is ill-defined and highly subjective, making comparisons difficult. Others warn it can be gamed or used to justify restrictive policies without robust verification.

See also AI alignment, ethics in artificial intelligence.