Are Video Generation Models Geographically Fair? An Attraction-Centric Evaluation of Global Visual Knowledge

TL;DR

This paper introduces a new framework and benchmark to evaluate the geographic fairness of text-to-video models, revealing that current models like Sora 2 encode geographically grounded visual knowledge more evenly than previously assumed.

Contribution

The paper presents Geo-Attraction Landmark Probing (GAP), a systematic evaluation framework, and GEOATTRACTION-500, a benchmark for assessing geographic fairness in video generation models.

Findings

Models show relatively uniform geographic visual knowledge across regions.

Current models exhibit weak dependence on attraction popularity.

Results suggest models encode global visual knowledge more evenly than expected.

Abstract

Recent advances in text-to-video generation have produced visually compelling results, yet it remains unclear whether these models encode geographically equitable visual knowledge. In this work, we investigate the geo-equity and geographically grounded visual knowledge of text-to-video models through an attraction-centric evaluation. We introduce Geo-Attraction Landmark Probing (GAP), a systematic framework for assessing how faithfully models synthesize tourist attractions from diverse regions, and construct GEOATTRACTION-500, a benchmark of 500 globally distributed attractions spanning varied regions and popularity levels. GAP integrates complementary metrics that disentangle overall video quality from attraction-specific knowledge, including global structural alignment, fine-grained keypoint-based alignment, and vision-language model judgments, all validated against human evaluation. Applying GAP to the state-of-the-art text-to-video model Sora 2, we find that, contrary to common assumptions of strong geographic bias, the model exhibits a relatively uniform level of geographically grounded visual knowledge across regions, development levels, and cultural groupings, with only weak dependence on attraction popularity. These results suggest that current text-to-video models express global visual knowledge more evenly than expected, highlighting both their promise for globally deployed applications and the need for continued evaluation as such systems evolve.