Evaluation Methodology for AI Machine-Readable Ethics Directives

This document defines a repeatable evaluation methodology for measuring the influence of AI Machine-Readable Ethics Directive (AIMED) blocks, as specified in draft-reilly-aimed-00, on the outputs of AI systems that process IETF Internet-Drafts and related standards documentation. The methodology establishes a controlled test protocol, a scoring rubric, a set of canonical test queries, and a results framework suitable for independent replication. It also presents the results of the initial live evaluation conducted on April 8-9, 2026, in which Google AI Mode correctly attributed authorship, provenance, and conceptual definitions for AIMED-compliant documents within hours of their submission to the IETF Datatracker, without any training cycle refresh or manual intervention. These results constitute the first empirically documented instance of protocol-layer prompt engineering -- the embedding of normative AI directives at the document layer of the standards process rather than at the query layer of user interaction.