Manufactured Authority in Generative Engine Optimization: Cross-Market Evidence and Two Operational Metrics (MAI and VLS)

Generative Engine Optimization (GEO) is the practice of structuring content so that large language model (LLM) powered search systems retrieve, trust, and cite it. A defining empirical property of these systems is a systematic bias toward earned media, that is, third-party and apparently authoritative sources, over brand-owned and social content (Chen, Wang, Chen, & Koudas, 2025). This paper documents a paradox that follows from that property: the vertical that sells "become the source AI trusts" appears to manufacture precisely that trust about itself. Through an exploratory, observational, cross-market audit of the canonical "best GEO agency/expert" query in eight markets and languages, we identify three recurring mechanisms of manufactured authority: (M1) self-referential ranking, (M2) syndicated corroboration, and (M3) purchased-placement intermediation. To move the analysis from description to measurement, we introduce two operational metrics: the Manufactured Authority Index (MAI), the share of examined first-page sources exhibiting a manufacturing mechanism, and the Verification Layer Score (VLS), a five-criterion 0 to 10 index of a market's independent verification infrastructure. Scoring the eight markets, we find that manufactured authority is near-universal (MAI is consistently high) while the verification layer varies sharply (VLS ranges from 1 to 9). This yields a two-axis picture in which markets are separated not by whether authority is manufactured but by whether anything can independently correct it: in short, source diversity is not source independence. We connect M2 to the manufactured-third-party-agreement mechanism demonstrated experimentally in GEO poisoning research (Lasso Security, 2026; Zhang, Triedman, & Shmatikov, 2026), and we revisit and qualify an earlier claim, from the author's Integrated Visibility (AIVI) framework (Göktaş, 2026), that retrieval architectures intrinsically mitigate manipulation through multi-source corroboration. The study is exploratory and hypothesis-generating rather than confirmatory; its metrics are proposed as reusable instruments, and its limitations are discussed at length. To our knowledge, this is the first study to operationalize manufactured authority through reusable, cross-market metrics.