Editor's Note: 

With the rapid development of AI tools such as DeepSeek,MKsports ChatGPT, and Doubao, these cutting-edge technologies are integrating with and influencing the protection, inheritance, and innovation of traditional Chinese culture in an unprecedented manner. These AI tools are not only capable of digitally processing a vast quantity of traditional cultural resources, but they also leverage 3D reconstruction technology to virtually repair and restore damaged cultural relics. This enables precious historical artifacts to be presented to the world in their complete form.

In this new series, the Global Times delves into the latest and most advanced AI tool applications in China, with an aim to explore how these tools can more effectively promote and preserve China's traditional culture and present it to the world.

An expert conducts restoration work on the murals at the Kizil caves in Northwest China's Xinjiang Uygur Autonomous Region. Photo: Courtesy of the Kizil Caves Research Institute

A groundbreaking initiative is underway to restore and revitalize the ancient murals of the Kizil Thousand-Buddha Caves in Northwest China's Xinjiang Uygur Autonomous ­Region. On Friday, the project was officially launched in Urumqi, the capital city of Xinjiang, aiming to use advanced technologies like AI to address the long-standing challenges of preserving these invaluable cultural relics.

"The significance of these efforts extends beyond mere restoration. The murals of the Kizil caves are a vivid record of the ancient Silk Road's prosperity and cultural diversity, making them an invaluable resource for studying the history, religion, and art of ancient Xinjiang," Li Qingming, director of the Kizil Caves Research Institute, told the Global Times.

Hi-tech empowered relics 

The Kizil caves, dating back to the late 3rd century, are among China's earliest and most extensive Buddhist cave complexes. These caves, with their richly detailed murals depicting the life of Sakyamuni, the founder of Buddhism, are not only a testament to ancient artistic brilliance, but are also a crucial link to the cultural exchanges along the ancient Silk Road. 

However, at the end of the 19th century and the beginning of the 20th century, Western expeditions plundered a large number of exquisite murals from the Kizil caves, resulting in the damage and loss of many mural sections. 

Meanwhile, centuries of natural erosion and human intervention have left many of these murals in a precarious state, with smoke damage and missing sections compromising their integrity, the Guangming Daily reported.

Zhao Li, a research fellow at the institute, who has dedicated over 30 years to mural conservation, highlighted the key challenges: Identifying and digitally restoring smoke-covered murals and accurately recognizing and completing missing sections using AI. 

She told the Global Times that, to tackle these issues, the project employs a combination of cutting-edge technologies. One team from the School of Art and Archaeology of Zhejiang University will use terahertz time-domain spectroscopy to non-invasively detect and identify the color layers of smoke-covered murals in Cave 161, enabling virtual restoration. 

Meanwhile, a digital group will leverage AI models and thought-chain strategies to complete the restoration of damaged murals in Cave 38. The whole project is scheduled to conclude in late May.

Zhao noted that while AI models excel at recognizing geometric patterns and architectural designs, they still require further training to accurately identify curved lines and human figures in the murals. However, the newly developed DeepSeek model, with its enhanced capabilities, offers hope for more precise and efficient restoration.

Wang Chaoyang, the chief architect of the digital culture laboratory for the project, told the Global Times that the project will follow a three-step approach: From technological breakthroughs to refining the tools and finally to cultivating a model for the restoration of damaged murals and the identification of smoke-stained murals. 

He said that as the project progresses, the hope is that the tools and techniques developed will not only benefit Xinjiang, but also be applicable globally, showcasing a "Xinjiang model" in cultural heritage preservation.

Way forward

The restoration of the Kizil caves is just one example of how technology, particularly AI, is revolutionizing the field of archaeology. Across the globe, archaeologists are leveraging AI to unlock the secrets of the past in ways that were once unimaginable.

For instance, AI-driven image recognition and data analysis are being used to identify and classify artifacts, predict archaeological sites, and even decipher ancient scripts, including the oracle bone script unearthed in the well-known Yinxu Ruins in Central China's Henan Province. 

In China, students are now taking courses like "AI Archaeology," where they learn to use AI to identify mold on paper relics or determine the age of bronze artifacts.

The potential is vast. Ruan Yongbin, an academician at the Chinese Academy of Sciences once pointed out, noting that among the vast number of materials collected by archaeologists, less than 10 percent are actually used for research. While AI can change this by efficiently processing and analyzing these underutilized resources.

Moreover, the trend toward interdisciplinary collaboration is gaining momentum. Universities in China are offering courses that combine archaeology with AI, preparing the next generation of scholars to bridge the gap between technology and cultural heritage.

As experts continue to explore the potential of AI in archaeology, the hope is that these technologies will not only restore ancient relics but also deepen our understanding of human history. As Li puts it, "the integration of technology and culture is not just about preserving the past; it's about ensuring that our cultural heritage continues to thrive in the future."