A proposed right to a human-comprehensible account of machine-mediated power: what happened, what was inferred, what authority applied and how to appeal and restore.
A decision explanation should answer what was observed, inferred, predicted and verified; what remains unknown; who had jurisdiction; what purpose allowed the use; what consequence is proposed; who received the data; how long it remains; and how appeal/restoration work.
A person may not need access to every model parameter to understand why institutional power was exercised. Legibility centers the accountable decision chain and evidence/authority basis.
Human notice, API record and audit log should share the same decision ID, authority provenance and epistemic types so that contestability is not lost between representations.
Source links establish traceability and support. They do not imply that the source endorses Concresca’s constitutional proposals.
Requires impact assessment, transparency, quality assurance, recourse, and graduated human involvement for covered federal automated decisions.
Government of Canada — official/source page ↗Defines a W3C Recommendation for representing provenance entities, activities, and agents.
W3C — official/source page ↗NO JUDGMENT WHATSOEVER. Concresca coordinates without assigning moral worth, character, guilt, danger, trustworthiness, loyalty, purity, normality, or social standing. Questions, thoughts, identities, messages, content, and conduct are not objects of Concresca judgment.