Does confidence drive influence in multi-agent deliberation systems?
When multiple AI agents deliberate together, does the agent who sounds most confident gain the most influence over the group's final answer? Understanding this matters because it determines whether consensus reflects actual competence or just persuasive miscalibration.
Multi-agent deliberation succeeds or fails not only on individual agents' predictions but on how they communicate and update. Modeling deliberation through Friedkin-Johnsen opinion dynamics — a tractable account of stubbornness, influence, and opinion change — yields a clean reframe: because the FJ parameters are input-dependent, deliberation behaves as a mixture of experts with adaptive routing. That explains when a multi-agent system beats single agents and static ensembles: when routing actually reflects agent competence on the input.
The problem is that competence is latent. In practice influence is established through observable proxies — an agent's self-assessed confidence, its perceived confidence, and its initial alignment with others. None of these is competence. So the routing that gives multi-agent systems their theoretical advantage is driven by the wrong signal, and miscalibrated confidence becomes influence. The paper names the resulting limitations precisely: miscalibrated agent confidence, misleading consensus, and routing errors.
This sharpens the vault's existing multi-agent failure cluster. Since Why do multi-agent LLM systems converge without genuine deliberation?, the FJ/MoE lens supplies the mechanism: agents route influence toward whoever sounds most confident, and consensus forms around persuasion rather than evidence — exactly the pattern When does debate actually improve reasoning accuracy? documents. The design implication is calibration-first: a multi-agent system is only as good as its agents' confidence is honest.
Inquiring lines that read this note 15
This note is a source for these research framings, grouped by the broader line of inquiry each explores. Scan the bold lines of inquiry; follow any specific question forward.
When and why does aggregating multiple reasoning samples improve performance?- Can calibrated confidence reduce misleading consensus in group deliberation?
- How does majority vote consensus handle cases where the consensus is confidently wrong?
- Does confidence-based weighting in deliberation substitute for competence-based expertise?
- Does agent influence correlate with competence or confidence in group reasoning?
- Does miscalibrated confidence in multi-agent deliberation create false consensus?
- Why does premature consensus form in multi-agent reasoning without genuine deliberation?
- What determines whether minority signals succeed in changing a group's consensus position?
- Why do some agent communities polarize while others reach consensus?
- Why might expanding group deliberation beyond five people produce bland consensus statements?
Related concepts in this collection 5
This note in its neighbourhood — explore the map, then jump to a related concept in the list below.
Click a node to walk · click center to open · click Open in graph to see this note in the full knowledge graph
-
Why do multi-agent LLM systems converge without genuine deliberation?
Multi-agent reasoning systems are designed to improve answers through debate, but often agents simply agree with early confident claims rather than genuinely disagreeing. What drives this pattern and how common is it?
the MoE/confidence-routing account explains why premature consensus forms
-
When does debate actually improve reasoning accuracy?
Multi-agent debate shows promise for reasoning tasks, but under what conditions does it help versus hurt? The research explores whether debate amplifies errors when evidence verification is missing.
confidence-as-routing-signal is the mechanism behind persuasion overriding evidence
-
When does adding more agents actually help systems?
Multi-agent systems often fail in practice, but the reasons remain unclear. This research investigates whether coordination overhead, task properties, or system architecture determine when agents improve or degrade performance.
both explain when MAS beats simpler ensembles; this one isolates routing quality as the condition
-
Can a quorum of validators really provide independent judgment?
If multiple validators share training data, prompts, evidence sources, or infrastructure, their agreement may reflect shared causes rather than independent confirmation. This could make quorum-based systems less reliable than they appear.
contrasts: a second route to a misleading consensus that needs no influence dynamics at all, only validators voting independently on the same shared inputs
-
Can true reports together mislead a group?
The SoK claims truthful reports can steer groups toward false beliefs, but provides no mechanism or citation. This explores whether honest inputs combined honestly can produce collective error, and what processes might explain it.
an open question that names this routing as one candidate mechanism (a confident true report outweighing better-grounded ones); the mapping is that note's reading, and this paper models opinion dynamics and not truthful reports
Related papers in this collection 8
Papers most semantically related to this note, ranked by cosine similarity in the embedding space.
- Multi-Agent Systems are Mixtures of Experts: Who Becomes an Influencer?
- ReConcile: Round-Table Conference Improves Reasoning via Consensus among Diverse LLMs
- Consensus is Strategically Insufficient: Reasoning-Trace Disagreement as a Knowledge-Representation Signal
- Finding Common Ground: Using Large Language Models to Detect Agreement in Multi-Agent Decision Conferences
- From Process Loss to Assembly Bonus: Human-Grounded Diagnosis of Multi-Agent LLM Collaboration
- A Framework for Studying AI Agent Behavior: Evidence from Consumer Choice Experiments
- Evaluating the Capabilities of LLMs for Persuasive Dialogue
- Large Language Models Cannot Self-Correct Reasoning Yet
Original note title
multi-agent deliberation is a mixture of experts whose routing tracks confidence not competence so miscalibration manufactures misleading consensus