The Inductive Interface: UI-Driven Human–AI Interaction under an Architecture of Attributed Responsibility

This working paper records a conceptual observation and architecture note within the CoAgnes / Harmogics research corpus. It specifies an interaction architecture for accountable human–AI work in domains such as regulatory review, compliance, audit, contemplative practice, and decision-support.

The paper develops the distinction between unfold and fold: unfold as the opening of a region of meaning into plurality, and fold as the non-delegable act of committing a statement attributable to a named witness. From this distinction it derives three connected contributions: the economic concept of fold debt, the interface pattern of an inductive interface, and the Omega Protocol of the Semantic Compass, where pages, JSON Schema, JSON Forms, agents, provenance, and responsibility-points form an inspectable architecture of attributed responsibility.

This deposit is published as a citable working artefact, not as an empirical study, legal advice, or completed software specification. It is intended as a foundation for later refinement, implementation-oriented prototypes, and possible development into a more formal paper.