# Uncovering Regional Defaults from Photorealistic Forests in Text-to-Image Generation with DALL-E 2

arXiv: https://arxiv.org/abs/2410.17255 · HTML: https://mlsanity.com/paper/2410.17255/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 12 references that are in the map.
- First posted on arXiv 3 Oct 2024.

## Builds on (most cited first, 7 of 7)

- [Language Models as Knowledge Bases?](https://mlsanity.com/paper/1909.01066/) (2019)
- [On the Opportunities and Challenges of Foundation Models for Geospatial Artificial Intelligence](https://mlsanity.com/paper/2304.06798/) (2023)
- [GPT4GEO: How a Language Model Sees the World's Geography](https://mlsanity.com/paper/2306.00020/) (2023)
- [AI's Regimes of Representation: A Community-centered Study of Text-to-Image Models in South Asia](https://mlsanity.com/paper/2305.11844/) (2023)
- [Philosophical Foundations of GeoAI: Exploring Sustainability, Diversity, and Bias in GeoAI and Spatial Data Science](https://mlsanity.com/paper/2304.06508/) (2023)
- [Measuring Geographic Diversity of Foundation Models with a Natural Language--based Geo-guessing Experiment on GPT-4](https://mlsanity.com/paper/2404.07612/) (2024)
- [Pre-Trained Language Models Represent Some Geographic Populations Better Than Others](https://mlsanity.com/paper/2403.11025/) (2024)

## Abstract (arXiv)

Regional defaults describe the emerging phenomenon that text-to-image (T2I) foundation models used in generative AI are prone to over-proportionally depicting certain geographic regions to the exclusion of others. In this work, we introduce a scalable evaluation for uncovering such regional defaults. The evaluation consists of region hierarchy--based image generation and cross-level similarity comparisons. We carry out an experiment by prompting DALL-E 2, a state-of-the-art T2I generation model capable of generating photorealistic images, to depict a forest. We select forest as an object class that displays regional variation and can be characterized using spatial statistics. For a region in the hierarchy, our experiment reveals the regional defaults implicit in DALL-E 2, along with their scale-dependent nature and spatial relationships. In addition, we discover that the implicit defaults do not necessarily correspond to the most widely forested regions in reality. Our findings underscore a need for further investigation into the geography of T2I generation and other forms of generative AI.