MultiUserWiki: Managing Editorial Divergence in Collaborative

LLM-Maintained Knowledge Bases

Bailing Zhang1,*

1 School of Computer Science and Data Engineering, NingboTech University, Ningbo 315336, China

Received: January 15, 2026 Revised: March 11, 2026 Accepted: May 15, 2026 ⋆ Corresponding author

ABSTRACT

When multiple LLM-assisted editors maintain a shared knowledge base, the conflicts that arise are seldom outright

factual contradictions. More often, two users reading the same source extract compatible assertions that differ in

specificity, coverage, or editorial emphasis. This paper asks a prior question: when should a conflict be resolved

by selecting a winner, and when should both perspectives be preserved? We study this question in a controlled

single-source collaborative setting. Five role-conditioned instances of GLM-4-Flash—simulating a senior researcher,

a teacher, a student, a clinician, and a conservative editor—independently extract assertions from AI-Wiki-21, a

21-page AI knowledge base. Their outputs yield 4,556 candidate conflict pairs; annotation of 138 selected pairs

produces a 100-conflict gold standard. Three findings emerge, each with scope limited to the single-source, single-

LLM-family, simulated-role setting studied. First, confirmed conflicts are exclusively editorial—54% Granularity

and 46% Coverage—with zero Factual conflicts, suggesting that in this controlled setting the dominant challenge is

editorial coordination, not truth arbitration. Second, direct LLM judgment achieves 97.5% agreement with human

annotators on this editorial-conflict gold standard, while a DPO-inspired quality heuristic (DIQS) reaches 81.0%

without API overhead; collective voting plateaus at 41.8%, echoing social-choice instability. Third, sensitivity

analysis of the fiber split threshold across τ ∈ [0.1,0.9] identifies τ∗ = 0.5 as the stable optimum, partitioning 3.1%

of conflicts while preserving diverse user perspectives. The results also raise ethical questions about automated

knowledge governance.

Keywords: Collaborative knowledge bases Editorial conflict resolution LLM-maintained wikis Preference aggregation

Knowledge governance Fiber structure

1. INTRODUCTION

Knowledge bases maintained by large language models

(LLMs) are increasingly envisioned as persistent, curated

repositories that multiple users contribute to and query—a

wiki-like structure in which the model acts as a disciplined

knowledge manager [1, 2]. Extending this to multi-user collaboration

introduces a governance challenge: when two

users produce different assertions about the same subject,

how should the knowledge base respond?

The intuitive answer—resolve the conflict by selecting the better

assertion—mischaracterises the problem. In collaborative

LLM-maintained knowledge bases, many apparent conflicts

are not logical contradictions. Two editors reading the same

source may extract one assertion at a high level of abstraction

and another at a finer grain; both can be simultaneously true

and mutually compatible. Treating such editorial divergence

as factual contradiction leads either to unnecessary arbitration

or to the silent discard of useful perspectives.

This paper studies the single-source collaborative setting: