AI models can compute certain urban planning tasks in seconds that would take humans 50 to 100 minutes to accomplish.
A recent study from Tsinghua University in China demonstrates that artificial intelligence (AI) and machine learning can create spatial layouts more efficiently than humans.
This conclusion stems from the goal of seeking solutions to improve cities in China that are suffering from congestion and over-concreting.
AI can perform certain steps in urban planning better than humans. (Illustration: Getty).
The researchers developed an AI system to handle the task of calculating numbers. This seemingly mundane work plays a crucial role in urban planning.
In fact, after implementing AI, the system was able to produce urban plans that surpassed human designs by about 50% across at least three metrics: accessibility to services, green space, and traffic flow.
Starting on a small scale, the research team tasked the AI with designing urban areas covering just a few square kilometers.
After two days of training and using several internal planning networks, the AI system was able to devise a logical layout for roads and land use for an entire city in just 15 minutes.
According to Yu Zheng, an automation scientist and the lead researcher, automation can save a significant amount of time in the planning process.
“AI models can calculate certain tasks in seconds that would take humans 50 to 100 minutes to perform,” Zheng shared.
The researcher noted that automating the most time-consuming urban planning tasks would allow managers to focus better on tasks such as resident placement and improving aesthetics.
Comparing the hybrid workflow process between AI and humans to a model based solely on human input, Zheng and colleagues found that the method could enhance accessibility by up to 12%.
He hopes that in the near future, the AI system will function as an assistant for urban planners, helping to generate algorithm-optimized ideas.
These ideas would still require expert involvement to review, adjust, and evaluate based on community feedback to create the most refined outcomes.
