Examining Agents' Bias Amplification versus Suppression in Multi-Agent Systems

2026-05-27

AI Summary

In multi-agent decision systems, individual biases can propagate and amplify, undermining overall fairness. This work addresses this issue by injecting group-preference biases into agents via prompt engineering and introduces a zero-centered metric—Favor Bias Strength (FBS)—that disentangles the effects of advantaged-group elevation from disadvantaged-group suppression. Experiments across diverse agent architectures, benchmark tasks, and state-of-the-art large language models reveal that systemic bias exhibits superadditive amplification even when input biases are uniformly distributed. Notably, when all agents share consistent biases, the emergent system-level bias exceeds the sum of individual biases, highlighting the severe risks posed by coordinated bias propagation in collaborative agent systems.

Abstract

Multi-agent systems are increasingly deployed to support various tasks where agents interact to achieve individual and collective objectives. Although these systems can enhance task performance and decision-making, fairness preservation through bias reduction remains challenging. This study examines how agent-level biases shift and impact system-wide fairness. We use prompts to expose individual agents to group-favoring bias, then assess downstream impacts at the system level. To quantify the impact, we propose Favor Bias Strength (FBS), a zero-centered metric that decomposes bias alteration between favored-group uplift and disfavored-group suppression. Using multiple agent designs, benchmarks, and up-to-date large language models, we show that agents endowed with bias can substantially affect system-wide fairness. Interestingly, when agents are exposed to bias uniformly, the system-wide bias elevates, even exceeding the additive sum of the individual agents' biases. The empirical evidence underscores the criticality of fairness in multi-agent systems, which warrants further analyses and empirical tests.

Problem

• bias amplification
• multi-agent systems
• fairness
• group-favoring bias
• system-wide bias

Innovation

• Favor Bias Strength
• bias amplification
• multi-agent systems
• fairness
• large language models