Aligned Agents, Biased Swarm: Measuring Bias Amplification in Multi-Agent Systems
2026-04-10 • Multiagent Systems
Multiagent SystemsArtificial Intelligence
AI summaryⓘ
The authors studied how groups of simple agents working together (multi-agent systems) can develop bias. They found that instead of reducing bias, complex workflows can actually make small random biases grow bigger, like an echo chamber. They built a test called Discrim-Eval-Open to measure bias by making agents compare different demographic groups. Their results showed that adding more complexity often made bias worse, even if each agent by itself was neutral. They also discovered that adding objective information sometimes made polarization happen faster.
Multi-Agent Systemsbias amplificationecho chambersfeedback loopsworkflow topologycomparative judgmentsethical robustnesssystemic polarizationbenchmark testingagent neutrality
Authors
Keyu Li, Jin Gao, Dequan Wang
Abstract
While Multi-Agent Systems (MAS) are increasingly deployed for complex workflows, their emergent properties-particularly the accumulation of bias-remain poorly understood. Because real-world MAS are too complex to analyze entirely, evaluating their ethical robustness requires first isolating their foundational mechanics. In this work, we conduct a baseline empirical study investigating how basic MAS topologies and feedback loops influence prejudice. Contrary to the assumption that multi-agent collaboration naturally dilutes bias, we hypothesize that structured workflows act as echo chambers, amplifying minor stochastic biases into systemic polarization. To evaluate this, we introduce Discrim-Eval-Open, an open-ended benchmark that bypasses individual model neutrality through forced comparative judgments across demographic groups. Analyzing bias cascades across various structures reveals that architectural sophistication frequently exacerbates bias rather than mitigating it. We observe systemic amplification even when isolated agents operate neutrally, and identify a 'Trigger Vulnerability' where injecting purely objective context drastically accelerates polarization. By stripping away advanced swarm complexity to study foundational dynamics, we establish a crucial baseline: structural complexity does not guarantee ethical robustness. Our code is available at https://github.com/weizhihao1/MAS-Bias.