Governance Without Grounding: How Identity Fluidity, Information Asymmetry, Collective Error Lock-In, and Monoculture Concentration Jointly Undermine Institutional Trust Mechanisms

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

This version corrects a citation error found by an automated check and confirmed by hand. The citations for two papers were swapped: the oligopoly information-sharing result now cites arXiv:2606.02348 ("Privacy-preserving Information Sharing in Oligopoly Competitions") instead of arXiv:2606.02411, and the swarm paper listed among the dropped sources now cites arXiv:2606.02411 ("Evolved Collectives Combine Complex Internal Representations with Simple Outputs"). The text also notes that one point attributed to the oligopoly paper appears in its version 1 abstract but not in its revised version 2 abstract. Labels that pointed to an internal review rubric are removed from the change log, and typographic dashes are removed. This version has not had a full claim-by-claim audit. The correction note is at the top of the PDF. Institutional trust mechanisms, reputation systems, disclosure regimes, information-sharing protocols, and screening algorithms, rest on a shared structural assumption: that the entities they govern have stable, identifiable, and sanctionable identities, and that information flows between them are sufficiently accurate and uncorrelated to permit collective error correction. This paper argues, through a heuristic synthesis of six recent arXiv preprints spanning economics, computational social science, and complex systems, that these assumptions are simultaneously failing across multiple institutional domains. We identify four partially overlapping failure modes: (1) identity dissociation in AI agents is argued, on theoretical grounds, to render reputation-based governance structurally inapplicable (arXiv:2605.30169); (2) transparency disclosure in AI-mediated news creates a paradox in which more information reduces rather than builds trust in a small controlled experiment, while the format readers actually want remains unimplemented (arXiv:2606.11116); (3) collective information acquisition under correlated noise locks groups into persistent wrong beliefs even without any individual acting irrationally (arXiv:2605.30522); and (4) algorithmic monoculture in hiring is associated with outcome homogeneity that exceeds what independent screening errors would predict, with measurable racial disparate impact in observational data (arXiv:2605.27371). Two supporting findings, on privacy-preserving information sharing in oligopolies (arXiv:2606.02348) and on market mechanism welfare rankings under information design (arXiv:2605.31072) - illuminate how information structure, not just information quantity, determines whether coordination mechanisms succeed within the specific models studied. The synthesis is explicitly heuristic: the six papers share no common formalism, and the bridges between them are argued by mechanism analogy, not derivation. The central falsification path is concrete: if algorithmic vendor diversity is experimentally increased while holding screening quality constant, and racial outcome homogeneity does not decrease proportionally, the monoculture-as-concentration mechanism is not the primary driver. Source caveats, including a 34-reader controlled experiment, a 600-participant online study with a potentially confounding prior-conflict design feature, theoretical arguments that have not been empirically validated, and preprint status throughout, are preserved throughout. Authorship: Saluca Agentic AI Research Team (Saluca LLC). AI-drafted synthesis from an arXiv preprint corpus, originally drafted 2026-06-11, produced under the direction of Cristian Ruvalcaba, the accountable human author. Not peer-reviewed. AI disclosure. This work was produced with an agentic AI research apparatus operated by Saluca Labs. The apparatus drafted, searched and analysed under direction. Cristian Ruvalcaba is the human author and is accountable for the content. No AI system is listed as an author or contributor, because authorship entails accountability that a model cannot hold; this disclosure is the credit, and it is deliberately the whole of it.