Democratizing City Planning with AI: New Research from Virginia Tech

**Virginia Tech researchers** are exploring innovative uses of _Large Language Models (LLMs)_ like **ChatGPT** and **Google's Gemini** to simplify city planning, making it more accessible for small to medium-sized cities. Traditionally, urban analysis required significant technical skills and was labor-intensive, often necessitating deep learning methods and manual image assessments. The new research, involving **Junghwan Kim**, an assistant professor and director of Smart Cities for Good at the university, discovered that LLMs could assess urban environments by analyzing street-view images, producing results similar to established deep learning approaches. This breakthrough democratizes access to advanced urban analytics, allowing stakeholders who may not possess technical skills to leverage AI for managing urban infrastructure more effectively. However, the researchers caution about potential biases in the AI's training data, which can lead to geographic disparities, often performing better in large cities than smaller towns. The work was published in _The Professional Geographer_ and collaborated with MIT's Kee Moon Jang. Despite challenges, such as AI hallucinations and assumption-making, the research marks a shift in how city planning might be conducted, promising a more inclusive and efficient future for urban management.