Exposing Blindspots: Cultural Bias Evaluation in Generative Image Models
This addresses cultural misrepresentation in generative AI for diverse users, providing a reproducible benchmark, though it is incremental by extending bias evaluation to image-to-image editing.
The study tackled cultural bias in generative image models by evaluating both text-to-image generation and image-to-image editing across six countries, revealing that models default to Global-North depictions, iterative editing erodes cultural fidelity, and superficial cues are applied. The results show culture-sensitive edits remain unreliable, with standardized data and protocols released for benchmarking.
Generative image models produce striking visuals yet often misrepresent culture. Prior work has examined cultural bias mainly in text-to-image (T2I) systems, leaving image-to-image (I2I) editors underexplored. We bridge this gap with a unified evaluation across six countries, an 8-category/36-subcategory schema, and era-aware prompts, auditing both T2I generation and I2I editing under a standardized protocol that yields comparable diagnostics. Using open models with fixed settings, we derive cross-country, cross-era, and cross-category evaluations. Our framework combines standard automatic metrics, a culture-aware retrieval-augmented VQA, and expert human judgments collected from native reviewers. To enable reproducibility, we release the complete image corpus, prompts, and configurations. Our study reveals three findings: (1) under country-agnostic prompts, models default to Global-North, modern-leaning depictions that flatten cross-country distinctions; (2) iterative I2I editing erodes cultural fidelity even when conventional metrics remain flat or improve; and (3) I2I models apply superficial cues (palette shifts, generic props) rather than era-consistent, context-aware changes, often retaining source identity for Global-South targets. These results highlight that culture-sensitive edits remain unreliable in current systems. By releasing standardized data, prompts, and human evaluation protocols, we provide a reproducible, culture-centered benchmark for diagnosing and tracking cultural bias in generative image models.