Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Debiasing International Attitudes: LLM Agents for Simulating US-China Perception Changes

Nicholas Sukiennik, Yichuan Xu, Yuqing Kan, Jinghua Piao, Yuwei Yan, Chen Gao, Yong Li

TL;DR

Results indicate that subjective news framing contributes only modestly to negative attitudes, whereas the devil's advocate agent proves most effective overall, suggesting that intermediate analytical steps can produce more human-like agent opinions.

Abstract

Large Language Models (LLMs) offer transformative opportunities to address the longstanding challenge of modeling opinion evolution in computational social science. This study investigates how media influences cross-border attitudes - a key driver of global polarization - by developing an LLM-agent framework to disentangle sources of bias and assess LLMs' capacity for human-like opinion formation in response to external information. We introduce an LLM-agent-based framework that models U.S. citizens' attitudes toward China from 2005 to 2025. Our approach integrates large-scale news data with social media profiles to initialize agent populations, which then undergo cognitive-aware reflection and opinion updating. We propose three debiasing mechanisms: (1) fact elicitation, extracting neutral events from subjectively framed news; (2) a devil's advocate agent that simulates critical contextualization; and (3) counterfactual exposure to surface inherent model biases. Simulations with two state-of-the-art LLMs (Qwen3-14b and GPT4o) reveal the expected negative attitudinal trend following media exposure. While all three mechanisms mitigate this trend to varying degrees, results indicate that subjective news framing contributes only modestly to negative attitudes, whereas the devil's advocate agent proves most effective overall, suggesting that intermediate analytical steps can produce more human-like agent opinions. Notably, the counterfactual study reveals contradictory findings across models, suggesting region-specific inherent biases tied to models' geographic origins. By advancing understanding of LLM-based opinion formation and debiasing methods, this study contributes to developing more objective models that better align with human cognitive tendencies.

Debiasing International Attitudes: LLM Agents for Simulating US-China Perception Changes

TL;DR

Results indicate that subjective news framing contributes only modestly to negative attitudes, whereas the devil's advocate agent proves most effective overall, suggesting that intermediate analytical steps can produce more human-like agent opinions.

Abstract

Large Language Models (LLMs) offer transformative opportunities to address the longstanding challenge of modeling opinion evolution in computational social science. This study investigates how media influences cross-border attitudes - a key driver of global polarization - by developing an LLM-agent framework to disentangle sources of bias and assess LLMs' capacity for human-like opinion formation in response to external information. We introduce an LLM-agent-based framework that models U.S. citizens' attitudes toward China from 2005 to 2025. Our approach integrates large-scale news data with social media profiles to initialize agent populations, which then undergo cognitive-aware reflection and opinion updating. We propose three debiasing mechanisms: (1) fact elicitation, extracting neutral events from subjectively framed news; (2) a devil's advocate agent that simulates critical contextualization; and (3) counterfactual exposure to surface inherent model biases. Simulations with two state-of-the-art LLMs (Qwen3-14b and GPT4o) reveal the expected negative attitudinal trend following media exposure. While all three mechanisms mitigate this trend to varying degrees, results indicate that subjective news framing contributes only modestly to negative attitudes, whereas the devil's advocate agent proves most effective overall, suggesting that intermediate analytical steps can produce more human-like agent opinions. Notably, the counterfactual study reveals contradictory findings across models, suggesting region-specific inherent biases tied to models' geographic origins. By advancing understanding of LLM-based opinion formation and debiasing methods, this study contributes to developing more objective models that better align with human cognitive tendencies.

Paper Structure

This paper contains 31 sections, 10 equations, 14 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Opinion Dynamics Modeling
  4. LLM-based Social Simulation
  5. Preliminary Data Collection
  6. Agent Initialization Data
  7. News Data Collection
  8. Ground Truth Attitudes
  9. Simulation Framework
  10. News Distribution
  11. Debiasing
  12. Reflection Mechanism
  13. Experiments and Results
  14. Overall Attitude Scores
  15. Attitude Trend Decomposition
  16. ...and 16 more sections

Figures (14)

  • Figure 1: The workflow for generating comprehensive agent profiles based on two real-world datasets.
  • Figure 2: Our framework for macro-scale cognitive-based attitudes evolution and debiasing.
  • Figure 3: Intervention mechanisms for news exposure.
  • Figure 4: Average score trends over time (left) and MAE comparisons (right) for GPT-4o and Qwen3 across four experiment types.
  • Figure 5: Positive/negative attitude rating trends (left) and MAE comparisons (right) for GPT-4o and Qwen3 across four experiment types
  • ...and 9 more figures