AI Use Statement
Generated by scripts/generate-traceability-report.py on 2026-06-28T10:31:47.
AI tools were used to assist with text preparation, candidate identification, organization of annotations, schema-shaped artifact drafting, validation support, and prose revision. All metaphor-identification decisions, conceptual mappings, interpretive claims, support scores, historical-corroboration claims, and final scholarly conclusions remain the responsibility of the researcher.
AI-generated suggestions are treated as provisional. They are not independent evidence, do not replace source inspection, and do not authorize claims without traceable support in corpus metadata, MIPVU decisions, CMT mappings, support ratings, historical notes, or validation artifacts.
The multi-model annotation stress test compares model behavior under bounded, blind instructions. Its outputs are diagnostic suggestions for human review. Model-to-model agreement and model-to-reference alignment are not human inter-annotator reliability, do not adjudicate accepted annotations, and do not prove that a scholarly interpretation is correct or reproducible.
Human reliability and adjudication, when present, are separate from AI assistance and model diagnostics. The package reports completed and partial states from validated human reliability status artifacts; absent or partial cohorts cannot be converted into publication-ready reliability claims.