Structural Fiscalistics in the AI Era ( WP 1.0)

Structural Fiscalistics in the AI Era proposes a pre-analytical framework for governing how “reality” enters AI and institutional systems. It addresses a central risk of generative models: coherent narratives produced without accountable institutional roots. The paper defines a simple, deployable pipeline: Fiscal Epistemology (FE) — decides what signals may enter (only rule-triggered, accountable, replayable records). Fiscal Geometry (FG) — renders those signals as structure (points, paths, bottlenecks), preserving gaps instead of filling them with narrative. Readouts — downstream metrics computed only from rendered structure. The design enforces one-way inheritance—evidence → structure → readout—so conclusions cannot fabricate facts. Missing evidence remains visible. The framework is domain-agnostic and demonstrated across tax, finance, education, and healthcare. It is intended as a governance interface for AI systems, regulatory platforms, and institutional analytics: a way to stabilize what counts as “real” before reasoning or scoring begins.