A groundbreaking study from mainland China has demonstrated how artificial intelligence (AI) can revolutionize archaeological research, specifically by analyzing ancient oracle bones to determine whether their authors were left or right-handed. The research team from Liverpool University in Suzhou and Renmin University in Beijing employed an unsupervised deep-learning tool called Bone2Vec to examine digital copies of these 3,000-year-old inscriptions.
The study aimed to explore whether humans’ right-handedness is rooted in biology or developed through social evolution. According to Wang Hefei, associate dean at Liverpool University International Business School and a co-author of the study, the findings suggest that handedness is likely a biological trait rather than one shaped by social factors. This discovery adds new depth to our understanding of human evolution and highlights the intricate role biology has played in shaping human behaviors over millennia.
The researchers analyzed over 1,000 images of oracle bone inscriptions, creating the largest dataset of its kind, scanned from artifacts stored at the Oracle Museum of China. By utilizing the Bone2Vec AI tool, the team was able to detect subtle patterns in the characters, which helped identify the dominant hand of the ancient scribes. These kinds of insights would be almost impossible to achieve manually, showcasing AI’s potential in uncovering hidden historical data.
The AI model, trained using artificial data to distinguish between left and right-handed writing, proved successful when applied to real oracle bones. Wang emphasized that high-quality data collection was essential for ensuring accurate AI-driven research results. As AI continues to be integrated into archaeological research, systematic methods of data collection will become increasingly critical.
Experts like Elic M. Weitzel, a postdoctoral research fellow at the Smithsonian National Museum of Natural History, noted that AI tools are already enhancing archaeological work, especially in artifact classification and pattern recognition. However, he cautioned that biases in training data could affect outcomes, reminding researchers of the importance of careful data preparation.
This study highlights not only the potential of AI in archaeology but also the necessity of interdisciplinary collaboration. Wang, with her background in computer science, partnered with archaeologists to explore the oracle bones, showcasing the future of academic research in integrating diverse fields and modern technologies.
READ MORE: