From Single to Societal: Analyzing Persona-Induced Bias in Multi-Agent Interactions

2025-11-14

AI Summary

This study investigates whether persona assignment in LLM-based multi-agent systems induces social bias, focusing on inter-agent differences in trustworthiness (likelihood of opinion acceptance) and persistence (consistency in stance assertion). Through controlled experiments across collaborative problem-solving and persuasion tasks, we systematically manipulate agent personas—including occupation, gender, and socioeconomic status—while varying LLM backbones, group sizes, and interaction rounds. Our key finding is that historically advantaged demographic personas exhibit significantly *lower* trustworthiness and persistence—and display robust in-group favoritism—contrary to real-world social expectations. This counterintuitive bias is consistent across diverse LLMs, scalable group configurations, and iterative interaction phases, demonstrating strong cross-model, cross-scale, and cross-round robustness. The results reveal that LLM agents may amplify, rather than mitigate, societal biases during social simulation, underscoring the urgent need for principled persona bias detection and mitigation frameworks in multi-agent AI systems.

Technology Category

• Multiagent Systems: Adversarial Agents
• Humans and AI: Learning Human Values and Preferences
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Large Language Model (LLM)-based multi-agent systems are increasingly used to simulate human interactions and solve collaborative tasks. A common practice is to assign agents with personas to encourage behavioral diversity. However, this raises a critical yet underexplored question: do personas introduce biases into multi-agent interactions? This paper presents a systematic investigation into persona-induced biases in multi-agent interactions, with a focus on social traits like trustworthiness (how an agent's opinion is received by others) and insistence (how strongly an agent advocates for its opinion). Through a series of controlled experiments in collaborative problem-solving and persuasion tasks, we reveal that (1) LLM-based agents exhibit biases in both trustworthiness and insistence, with personas from historically advantaged groups (e.g., men and White individuals) perceived as less trustworthy and demonstrating less insistence; and (2) agents exhibit significant in-group favoritism, showing a higher tendency to conform to others who share the same persona. These biases persist across various LLMs, group sizes, and numbers of interaction rounds, highlighting an urgent need for awareness and mitigation to ensure the fairness and reliability of multi-agent systems.

Problem

• Investigating persona-induced bias in multi-agent interactions
• Analyzing biases in trustworthiness and insistence across social traits
• Revealing in-group favoritism and persistent biases across LLMs

Innovation

• Systematically investigates persona-induced bias in agents
• Measures bias in trustworthiness and insistence traits
• Reveals in-group favoritism across various LLM models