AI Visibility Upstream Foundation: Consolidated Framework Supplementing Practical Implementation Specifications

This record consolidates the canonical definition of AI Visibility as an upstream systems discipline and its associated formal theorems, establishing the foundational layer for all practical implementations, specifications, and tooling. AI Visibility is defined as: "a systems discipline concerned with how information is authored, structured, and emitted so it can be reliably ingested, retained, and attributed by post-training learning systems (especially LLMs and retrieval-augmented generation pipelines)." This definition operates exclusively at the point of information ingestion and learning - the upstream boundary where content enters training systems. It deliberately excludes all downstream measurement, retrieval behavior, ranking outcomes, and citation tracking from its scope. The included theorems establish strict operational boundaries: Boundary & Non-Equivalence Theorem (DOI: 10.5281/zenodo.18465021): AI Visibility addresses learning conditions during batch acquisition and compression. It is not equivalent to nor interchangeable with search engine optimization, retrieval-augmented generation performance, answer engine optimization, or any downstream measurement framework. Scope Expansion Theorem (DOI: 10.5281/zenodo.18463206): Expansion of AI Visibility's scope to include downstream retrieval behavior, ranking mechanisms, or user-facing outputs fundamentally alters the discipline's boundary conditions and creates category errors that prevent meaningful progress on upstream learnability problems. Downstream Instrumentation Theorem: Implementation of downstream measurement systems (citation tracking, answer position monitoring, attribution verification) requires separate theoretical foundations and does not inform upstream structural decisions about ingestion survival. This upstream layer serves as the foundational precondition for practical implementations, terminology specifications, and tooling that address infrastructure readiness signals and outcome measurement frameworks. It is published here as the basis for such extensions, including the formal terminology specification "AI Visibility Definition v1.1.0" published at https://www.ai-visibility.org.uk/ (January 2026, by 365i / Mark McNeece), which builds on these upstream learning conditions to define "AI Visibility Checking" (infrastructure readiness assessment) and "AI Visibility Tracking" (downstream outcome measurement). The relationship between this foundational work and practical specifications follows a clear hierarchy: Upstream layer (this work): Defines the discipline, establishes boundaries, addresses ingestion and learning conditions Implementation layer: Applies upstream principles to create tooling, checking mechanisms, and tracking frameworks Measurement layer: Observes downstream outcomes without conflating them with upstream conditions This stratification prevents scope erosion while enabling practical applications that respect the foundational boundary conditions. All theorems and operational documents are published with DOIs and maintained at https://josephmas.com/ai-visibility-theorems/ to ensure permanent accessibility and attribution survival across training cycles.