The Urban Planning Society of Shenzhen (UPSSZ) convened the "AI Urban Design Showcase: Emerging Practices in the Greater Bay Area" salon on August 9th at Shenzhen X- Institute, drawing 120+ planning professionals from Hong Kong, Guangzhou, Shenzhen, and Zhuhai for intensive cross-border exchanges on AI-driven urban innovation.

Co-organized by the Greater Bay Area Urban Designers Professional Alliance and hosted by UPSSZ, the salon featured technical presentations, live demonstrations, and strategic dialogues centered on integrating AI into urban design workflows.

UPSSZ Vice President Huang Weidong set the tone: "We stand at an inflection point where AI redefines urban design's boundaries. The challenge lies not just in adopting technology, but in harnessing its transformative power to create livable, resilient cities."

Technical Breakthroughs Showcased

AIGC Image Tech: Revolutionizing Planning Workflows

He Songlun (Urban Planning & Design Institute of Shenzhen ) elaborated on how the institute leverages the open-source framework Stable Diffusion for customized training and development of diffusion models, aiming to enhance the practical effectiveness and efficiency of AIGC technology in the field of urban planning and design. This approach achieves a harmonious blend of precision and expressiveness in generating large-scale aerial perspective renderings.

AI Optimization Methodologies in Urban Design

Dr. Zheng Hao (City University of Hong Kong) presented breakthroughs achieved by his team in the field of AI-assisted urban design. He positioned the computer as a design collaborator, integrating big data on urban architectural layouts, crime rates, pedestrian activity levels, and surface temperatures as AI training sets to achieve intelligent generation and optimization of urban layouts. The AI model can further extend its evaluation functions to precisely analyze urban characteristics such as pedestrian accessibility, surface temperature regulation effects, and pandemic transmission risks, providing a scientific basis for constructing safer and more livable urban environments. This innovation not only improves design efficiency but also enhances the scientific and predictive nature of urban planning.

Cross-Domain Traffic Simulation & Intelligent Control

Wu Ruoqian (Shenzhen Transport Planning Center) believes that while traffic simulation technology has a history spanning decades, it is undergoing new transformations due to changes in demand for refined and intelligent management, as well as the application of AI and large models. China's urban transportation infrastructure is already globally leading, but with a slowdown in investment scale, the current focus has shifted to optimizing existing resources. Future transportation systems will be more autonomous, with reduced human involvement, especially driven by the development of autonomous vehicles and drones. Shenzhen Transport Planning Center Co., Ltd. has developed TransSim, a fully autonomous and controllable medium- and micro-scale traffic simulation software, demonstrating intelligent expansion application cases for expressways, key corridors, important districts, rail networks, and key hubs based on AI traffic agent construction and simulation technologies.

Controllable LLM Strategies for Urban Planning Software

Dr. Huang Duo (South China University of Technology) ’s sharing was mainly divided into three parts. The first part introduced the latest progress and key technical features of large language models. The second part elaborated on how to construct application strategies for controllable large language models from perspectives such as technical vision, overall strategy, and multi-dimensional technical strategies, based on discussing the dilemmas faced by urban planning. The third part discussed the development of a large language model software system that integrates technologies such as NLP, RAG, PE, Agent, and LLM generation, based on the establishment of a local knowledge base, targeting functions such as planning text archive management, semantic vector library management, comprehensive retrieval, basic and advanced intelligent language model applications, and comprehensive text generation.

AI-Social Intelligence Synergy: Future Scenarios

Dr. Yuan Xiaohui (Tencent Research Institute) proposed that we are at a node of intellectual transformation, where intellectual resources have become a service accessible to everyone. Against this backdrop, he introduced Tencent's attempts and efforts in relying on urban spaces to enhance social computing power and promote technological innovation, proposing that urban spaces are the value carriers in the AI era and introducing the concepts of social intelligence and social computing power. Based on the viewpoint that "the evolution of urban form is also a process of continuously enhancing social computing power," he presented three possible future scenarios for human society: AI empowering various industries, comprehensive employment replacement, and moving towards human-machine symbiosis.

AI-Driven Collaborative Urban Innovation Paradigms

Liu Long (LAY-OUT Planning Consultants Co., Ltd) taking six product projects of LAY-OUT as examples, Liu Long explored the practical R&D work of achieving "four reshapings" in AI+ industries and "everyone creating beautiful cities" in AI+ cities through AI technology enhanced by professional knowledge and vertical models in the planning industry. He proposed moving beyond the ivory tower of the industry, facing the public, and using action-initiated planning to cope with the active growth of urban value, meet the needs of refined urban management and planning and design, and promote the digital transformation of planning and design.

Expert Critiques

Dr. Yuan Xiaohui, commented on the first two presentations, stating, "He Songlun demonstrated how AIGC can be applied to generate renderings in planning and design to assist practical work, providing a best practice case that has been successfully implemented. She also explored the role relationship between AI and designers, which holds significant implications for future promotion and replication. Dr. Zheng Hao's presentation showcased a methodology based on urban data, transforming interesting propositions into quantifiable and assessable scientific questions, helping us re-understand cities and discover patterns. The combination of these two approaches provides valuable insights for the application of AI and science in urban research."

Alain Chiaradia, Honorary Associate Professor at the University of Hong Kong and Deputy Director of the Department of Urban Planning and Design at the University of Hong Kong, commented on the presentations by Wu Ruoqian and Dr. Huang Duo. "The above two excellent presentations offer starkly different perspectives. The first speech focuses on transportation, showcasing concepts such as transportation service levels, the value of time, and free flow. In the past, people considered transportation a waste of time, but now with technological advancements, the value of time spent traveling has become more complex. The next challenge lies in enhancing journey quality, not just shortening travel time. The second speech explores the holistic and complex nature of urban planning and design, emphasizing the importance of using technologies like AI to address planning challenges. I would also like to introduce a new challenge for everyone to ponder: How do we rethink urban planning from a people-centric perspective in the future? Do traditional planning and design methods need to be overturned? How do we integrate the experiences of urban residents into planning and design?"

Zhao Pengjun (Academician of the Academy of Social Sciences in the UK, Dean and Professor of the School of Urban Planning and Design at Peking University, and Vice President of the Urban Planning Society of Shenzhen) commented, "The urban planning industry is at a critical period of transformation, with the introduction of AI serving as an important driving force for this transformation. However, the underlying power source (spontaneous innovation or passive adaptation) behind this transformation still needs to be clearly defined. Meanwhile, industry standardization and replicability have become urgent issues to be addressed. The existing standards and technologies have limited capabilities for popularization and replication, and there is an urgent need to establish unified standards and promote scenario integration. In response to the challenges brought about by new technologies, such as the digital divide and the effectiveness of technology guidance, the industry has generally recognized the importance of unified standards. To this end, all parties should strengthen cooperation, implement precise measures, promote the formulation and implementation of industry standards, ensure the efficient and widespread application of AI technology in urban planning, and jointly promote the healthy development of this field."

During the open exchange session, the guests took turns sharing their insights and expectations regarding the development of AI and the urban planning and design industry. The audience actively posed questions and exchanged ideas with the guests on topics such as the evaluation mechanism for AI, the integration of AI in education, and the application of AI in the planning and design of Dachan Bay in Shenzhen, creating a lively atmosphere at the venue.

At the close of the event, Deputy Secretary-General Sun Yuehua highlighted that this year marks the 30th anniversary of the Shenzhen Urban Planning Society, and this event serves as a remarkable part of the society's series of academic celebrations. By the end of this year, Shenzhen, as the rotating host city of the Greater Bay Area Urban Designers Alliance for the 2023-2024 term, will organize the Greater Bay Area International Urban Design Conference and plan more diverse academic exchange activities. Stay tuned for more exciting developments.