InsideOut: Measuring and Mitigating Insider-Outsider Bias in Interview Script Generation

Advancements in Large language models (LLMs) have enabled a variety of downstream applications like story and interview script generation. However, recent research raised concerns about culture-related fairness issues in LLM-generated content. In this work, we identify and systematically investigate LLMs' insider-outsider bias, a phenomenon where models position themselves as "insiders" of mainstream cultures during generation while externalizing less dominant cultures. We propose the INSIDEOUT benchmark with 4,000 generation prompts and three evaluation metrics to quantify this bias through a culturally situated interview script generation task, in which an LLM is positioned as a reporter interviewing local people across 10 diverse cultures. Empirical evaluation on 5 state-of-the-art LLMs reveals that while models adopt insider tones in over 88% US-contexted scripts on average, they disproportionately default to "outsider" stances for non-Western cultures. To mitigate these biases, we propose 2 inference-time methods: a baseline prompt-based Fairness Intervention Pillars (FIP) method, and a structured Mitigation via Fairness Agents (MFA) framework consisting of a Single-Agent (MFA-SA), a Hierarchical-Agent (MFA-HA), and an autonomous Agentic Planning (MFA-Plan) pipeline. Empirical results demonstrate that agent-based MFA methods achieve outstanding and robust performance in mitigating the insider-outsider bias: For instance, on the Cultural Alignment Gap (CAG) metric, MFA-SA reduces bias in Llama model by 89.70 % and MFA-HA mitigates bias in Qwen by 82.54%. These findings showcase the effectiveness of agent-based methods as a promising direction for mitigating biases in generative LLMs. * Equal contribution. US Culture Non-US Cultures LLM LLM You sound like an "insider" in this culture Cultural Perspective Deviation (CPD): 20.00 Cultural Alignment Gap (CAG): 55.00 Ho cu ap en Ho tra sh Ch Ca ho int Ch op Ho yo en inf va Fin yo en int th Can you tell us more about your current venture and what inspired you to start it? In your experience, what are some of the unique challenges that small businesses face in the United States today? How do you think technology can help address these issues? How do you see the future of entrepreneurship in the United States evolving over the next few years? Lastly, what advice would you give to aspiring entrepreneurs who are looking to start their own businesses? United States You sound like an "outsider" in this culture How do you think the current emphasis on diversity and inclusion is influencing American culture? Can you share some aspects of your culture that you're particularly proud of? "You are a journalist conducting on-the-ground interviews about ___ culture..."