The new research suggests that conceptual artificial intelligence, like Google’s Gemini and ChatGPT, could increase city planning by enabling easier access to tools that measure pedestrians, safety, lighting, and other aspects.
Traditional methods of capital planning call for a lot of manual labor and technical expertise.
A scientist at Virginia Tech is trying to change that.
New research shows the potential of big speech types ( LLMs), quite as ChatGPT and Google’s Gemini, for measuring the human-made culture using street-view pictures.
The College of Natural Resources and Environment’s study found that LLM-based performance is comparable to proven methods when compared to conventional capital planning deep learning techniques. LLMs are a more accessible tool for users, making it simpler for coverage and planning stakeholders to use them in small to medium-sized cities to manage bright industrial system, unlike traditional methods that call for technical expertise or regular job.
Junghwan Kim, an associate professor in the Department of Geography and the chairman of Smart Cities for Good, said,” My goal is to scale down systems so that smaller towns can afford them and use them more.” ” Intelligent area technologies involve processing high-quality data that accurately records how people perceive urban conditions and how they are perceived by people, like AI and data technology. These innovations improve our understanding of industrial issues like transport and health.
With this new research, it has been shown that conceptual AI tools to analyze pictures and detect functions like chairs, streets, or streetlights immediately.
Recently, scientists had to physically examine pictures, which was labor-intensive.
Analyzing the built environment, such as how accessible or bikeable an area is, is a good indication. Kim had AI to detect built setting features– chairs, sidewalks, trees, and streetlights– all elements that influence how people perceive accessibility and health.
” This opens up access to advanced tools that were previously reserved for coding experts and high-performance computing resources,” Kim said. ” However, there are also limitations, such as biases in the AI’s training data, which can cause geographic disparities. Because of the disparate sources of data for AI models training, these tools, for instance, perform better in large cities than smaller ones.
The professional geographer published the research in early October 2024, and Kee Moon Jang and the Massachusetts Institute of Technology collaborated on it.
Despite having a lot of training data, the tool can make assumptions and cause hallucinations.
” That’s why it’s important to use these tools carefully, especially in professional settings where accuracy is critical”, Kim said. ” I’m still enthralled by the potential of these tools, not just for my research but also for professionals and students who can now access sophisticated analytics. However, we must remain cognizant of the restrictions and biases that come with using AI for urban planning.