When LLMs Imagine People: A Human-Centered Persona Brainstorm Audit for Bias and Fairness in Creative Applications

TL;DR

This paper presents the Persona Brainstorm Audit (PBA), a scalable method for detecting biases in open-ended persona generation by LLMs, revealing complex, intersectional, and nonlinear bias patterns across models.

Contribution

Introduces PBA, a novel bias detection method for creative LLM outputs that accounts for intersectionality and provides interpretable severity labels.

Findings

Bias varies nonlinearly across model generations.

Larger models are not always fairer.

Intersectional biases can be hidden in single-axis metrics.

Abstract

Large Language Models (LLMs) used in creative workflows can reinforce stereotypes and perpetuate inequities, making fairness auditing essential. Existing methods rely on constrained tasks and fixed benchmarks, leaving open-ended creative outputs unexamined. We introduce the Persona Brainstorm Audit (PBA), a scalable and easy to extend auditing method for bias detection across multiple intersecting identity and social roles in open-ended persona generation. PBA quantifies bias using degree-of-freedom-aware normalized Cram\'er's V, producing interpretable severity labels that enable fair comparison across models and dimensions. Applying PBA to 12 LLMs (120,000 personas, 16 bias dimensions), we find that bias evolves nonlinearly across model generations: larger and newer models are not consistently fairer, and biases that initially decrease can resurface in later releases. Intersectional analysis reveals disparities hidden by single-axis metrics, where dimensions appearing fair individually can exhibit high bias in combination. Robustness analyses show PBA remains stable under varying sample sizes, role-playing prompts, and debiasing prompts, establishing its reliability for fairness auditing in LLMs.