Generative AI-driven architectural and structural engineering designs: A review

The fields of architectural and structural design are undergoing a paradigm shift driven by breakthroughs in generative artificial intelligence (AI) technologies. Advanced generative AI, such as generative adversarial networks (GANs), variational autoencoders (VAEs), diffusion models, and multimodal large models, can learn from existing design data to achieve automated generation of solutions, thereby considerably expanding the creativity and efficiency boundaries of design. In the field of architectural design, GANs and VAEs have been widely applied to generate innovative forms and spatial layouts. In addition, state-of-the-art diffusion models and three-dimensional (3D) generative models are being integrated into design workflows. The field of structural design is also experiencing a revolution, where generative AIs (e.g., GANs, VAEs, and diffusion models) are gradually achieving intelligent design of structural component layouts and dimensions. Furthermore, a fundamental barrier persists in generative AI due to mismatches in data representation and design objectives. Consequently, research on integrated architectural–structural design remains limited, underscoring multi-disciplinary collaborative AI as a promising avenue for future investigation. In this study, the current state of research on generative AI in architectural and structural designs is systematically reviewed. The future directions and challenges of mainstream research are analyzed. This review can serve as a reference for the development of generative AI-driven architectural and structural designs.