Generative AI and Chinese Artistic Fonts: Cultural Preservation, Digital Creativity, and Educational Perspectives

Abstract

Generative artificial intelligence has increasingly influenced the preservation and transformation of traditional artistic practices in the digital era. Among these technologies, Generative Adversarial Networks (GANs) have demonstrated strong capabilities in the generation of Chinese artistic fonts and calligraphic styles. This paper examines the role of GAN-based frameworks in supporting the preservation, reconstruction, and creative transformation of Chinese artistic typography from cultural, social, and educational perspectives. The study reviews representative GAN architectures, including DCGAN, Pix2Pix, CycleGAN, StyleGAN, MobileGAN, and GigaGAN, together with specialised artistic font generation models such as zi2zi, S2PNet, TET-GAN, F2PNet, and StrokeGAN. Beyond technical performance, the paper analyses how these technologies contribute to cultural heritage preservation, digital creativity, typography education, and the accessibility of artistic production. The discussion further examines concerns regarding artistic authenticity, cultural identity, intellectual ownership, and the changing relationship between human creativity and AI-generated art. The findings indicate that GAN-based systems have significant potential in preserving endangered artistic traditions and supporting digital humanities research. At the same time, challenges remain regarding model controllability, evaluation standards, ethical considerations, and dependence on training data. The paper concludes by highlighting future research opportunities involving multimodal generative systems, educational applications of AI-generated typography, and internationally transferable frameworks for preserving artistic