# Towards Automated Urban Planning: When Generative and ChatGPT-like AI Meets Urban Planning

**Authors:** Dongjie Wang, Chang-Tien Lu, Yanjie Fu  
**arXiv:** [2304.03892](https://arxiv.org/abs/2304.03892)  
**Date:** 2023-04-08  
**Summary page:** https://www.summarizepaper.com/en/paper/2304.03892v1/

## Key points

- Urban planning and artificial intelligence (AI) are two separate fields that can benefit from cross-pollination.
- Urban planning is important for sustainability, living conditions, economic development, disaster management, and environmental conservation.
- AI can contribute to modern urban planning through adversarial learning, generative neural networks, deep encoder-decoder networks, conversational AI, geospatial and temporal machine learning.
- Automated land-use configuration is a central problem discussed in the paper which involves generating land uses and building configurations for a target area using surrounding geospatial data along with human mobility patterns, social media activity data, environmental factors like air quality or noise levels as well as economic activities.
- Implications of AI for urban planning include optimizing resource allocation in cities while also improving quality of life for residents and exploring how chatbots powered by natural language processing (NLP) techniques could help citizens engage with city planners more effectively.
- Cross-disciplinary collaboration between urban planning and AI is crucial for future research in this area.

## Summary

In their paper titled "Towards Automated Urban Planning: When Generative and ChatGPT-like AI Meets Urban Planning," Dongjie Wang, Chang-Tien Lu, and Yanjie Fu explore the intersection of urban planning and artificial intelligence (AI). They highlight that while these two fields have developed separately, there is now increasing interest in cross-pollination to benefit from each other's advances. The paper emphasizes the importance of urban planning from various perspectives such as sustainability, living conditions, economic development, disaster management, and environmental conservation. The authors review fundamental concepts of urban planning and relate them to crucial open problems of machine learning. They discuss adversarial learning, generative neural networks, deep encoder-decoder networks, conversational AI, geospatial and temporal machine learning to demonstrate how AI can contribute to modern urban planning. One central problem discussed is automated land-use configuration which involves generating land uses and building configurations for a target area using surrounding geospatial data along with human mobility patterns, social media activity data, environmental factors like air quality or noise levels as well as economic activities. The authors conclude by delineating some implications of AI for urban planning and proposing key research areas at the intersection of both topics. They suggest that automated land-use configuration could be used to optimize resource allocation in cities while also improving quality of life for residents. Additionally they propose exploring how chatbots powered by natural language processing (NLP) techniques could help citizens engage with city planners more effectively. Overall this paper provides a comprehensive overview of how AI can contribute to modern urban planning practices. It highlights the potential benefits that can be achieved through cross-disciplinary collaboration between these two fields and proposes several avenues for future research in this area.