A new study led by Virginia Tech highlights the potential of AI tools like ChatGPT in city planning, promising greater accessibility and efficiency for small to medium-sized cities.
Traditional city planning methods often require significant technical expertise and extensive manual labor. However, new research led by Virginia Tech promises to transform how cities, especially small to medium-sized ones, manage their urban landscapes.
The study examined the potential of large language models (LLMs), such as ChatGPT and Google's Gemini, in assessing human-made environments using street-view images.
By comparing the performance of LLMs to traditional deep learning methods used in city planning, the researchers found that LLM-based approaches are similarly effective.
"My goal is to scale down technologies, making them more affordable and effective for smaller cities," corresponding author Junghwan Kim, an assistant professor in the Department of Geography at Virginia Tech and the director of Smart Cities for Good, said in a news release. "Smart city technologies involve using advanced urban analytics, like AI and data science, to process high-quality data that captures urban environments and how people perceive them. These technologies help us better understand urban issues, such as transportation and health."
The key advantage of using LLMs lies in their accessibility. Unlike conventional methods that demand technical skills and manual efforts, LLMs allow policymakers and planners to efficiently analyze urban features. This includes identifying elements such as benches, sidewalks, trees and streetlights that influence an area's walkability and safety perceptions.
"This democratizes access to advanced tools that were once only usable by experts with coding skills and high-performance computing resources," Kim added.
However, he acknowledged the AI's limitations, adding, "there are also limitations, such as biases in the AI's training data, which can cause geographic disparities. For example, these tools tend to perform better in large cities than in smaller towns because of the uneven availability of data for training the AI models."
The research, conducted in collaboration with Kee Moon Jang, a postdoctoral researcher at MIT Senseable City Lab, was recently published in The Professional Geographer. It demonstrates that AI can automatically detect built-environment features from images, significantly reducing the labor involved in manual analysis.
Despite its promise, AI's application in urban planning is not without its challenges.
"That's why it's important to use these tools carefully, especially in professional settings where accuracy is critical," Kim added. "I'm still excited about the potential of these tools, not only for my research but also for students and professionals who can now easily access advanced analytics. However, we must remain aware of the limitations and biases that come with using AI in urban planning."
This study helps set the stage for a future where city planning is not only more efficient but also more inclusive, enabling even the smallest cities to leverage cutting-edge technologies for their development.