Consensus is Strategically Insufficient: Reasoning-Trace Disagreement as a Knowledge-Representation Signal
2026-06-02 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors explain that in systems with many agents, like AI tools working together, disagreement isn't always a mistake—it can mean they see things differently because of uncertain values. They create a way to label these disagreements based on how agents think and decide, grouping them into four types. This helps the system decide how to handle disagreements better. They show how this works for online content checks and suggest it mixes clear logic with complex AI reasoning effectively.
multi-agent systemsdisagreement statesreasoning tracescontent moderationsymbolic knowledge representationlarge language modelsstrategic routingnormative uncertainty
Authors
Michał Wawer, Jarosław A. Chudziak
Abstract
Multi-agent systems are commonly designed to reduce disagreement through voting, consensus protocols, debate, or fault-tolerant aggregation. We argue that this objective is insufficient for value-laden tasks, where disagreement may reflect genuine normative uncertainty rather than agent error. Building on prior work on reasoning-trace disagreement in human-AI collaborative moderation, we propose a knowledge-representation layer in which reasoning traces and agent decisions are abstracted into symbolic disagreement states. Given agents producing explicit reasoning traces and binary decisions, we distinguish four states according to reasoning similarity and conclusion agreement: convergent agreement, divergent agreement, convergent disagreement and divergent disagreement. These states support defeasible strategic routing rules. We instantiate the framework in content moderation and argue that disagreement-aware routing provides a bridge between sub-symbolic LLM deliberation and symbolic knowledge representation for multi-agent strategic reasoning.