Conformity Phase-Transition Bias Lock-In

Properties

family: lock-in
severity: systemic
triggerError: Debating agents each carry a small pre-existing individual bias, and the debate is run under a given sampling noise level (e.g. LLM temperature).
mechanism: Using a physics-inspired social-dynamics model, the paper shows debate outcomes undergo a phase transition to collective bias once conformity pressure among agents surpasses a critical threshold set jointly by the agents' initial bias and the debate's noise level; below the threshold, individual biases stay individual, but above it they lock in as a shared, amplified group norm. Agent heterogeneity smooths (rounds off) the sharpness of this transition rather than preventing it.
blastRadius: Any deployment that treats multi-agent debate consensus as a de-biased signal is exposed once the conformity threshold is crossed; the authors show the effect generalizes beyond toy settings to realistic decision tasks including investment decisions and LLM-as-a-judge evaluation.
mast: [Cat 2][Cat 3]
detection: Track finite-size crossover behavior in measured bias as conformity, initial bias, and sampling noise are varied, the paper's own experiments confirm the predicted phase transition this way; watch for a sharp jump in group-level bias as temperature or agent homogeneity changes rather than assuming bias grows only gradually.
mitigation: Increase agent heterogeneity (different models, prompts, or initializations) in the debate pool, which the paper shows smooths the phase transition and suppresses emergence of collective bias; keep debate noise and conformity pressure below the derived critical threshold.
source: Emergence of Biased Consensus in Multi-Agent LLM Debates (arXiv:2608.02827)

Provenance

Source: https://arxiv.org/abs/2608.02827

Properties

family

lock-in

severity

systemic

triggerError

Debating agents each carry a small pre-existing individual bias, and the debate is run under a given sampling noise level (e.g. LLM temperature).

mechanism

Using a physics-inspired social-dynamics model, the paper shows debate outcomes undergo a phase transition to collective bias once conformity pressure among agents surpasses a critical threshold set jointly by the agents' initial bias and the debate's noise level; below the threshold, individual biases stay individual, but above it they lock in as a shared, amplified group norm. Agent heterogeneity smooths (rounds off) the sharpness of this transition rather than preventing it.

blastRadius

Any deployment that treats multi-agent debate consensus as a de-biased signal is exposed once the conformity threshold is crossed; the authors show the effect generalizes beyond toy settings to realistic decision tasks including investment decisions and LLM-as-a-judge evaluation.

mast

[Cat 2][Cat 3]

detection

Track finite-size crossover behavior in measured bias as conformity, initial bias, and sampling noise are varied, the paper's own experiments confirm the predicted phase transition this way; watch for a sharp jump in group-level bias as temperature or agent homogeneity changes rather than assuming bias grows only gradually.

mitigation

Increase agent heterogeneity (different models, prompts, or initializations) in the debate pool, which the paper shows smooths the phase transition and suppresses emergence of collective bias; keep debate noise and conformity pressure below the derived critical threshold.

source

Emergence of Biased Consensus in Multi-Agent LLM Debates (arXiv:2608.02827)