For more than a century, the giant designs etched into the southern Peruvian desert known as the Nazca Lines — depicting animals, plants, geometric shapes, and human figures — have baffled researchers. These enormous geoglyphs, created by the Nazca civilization between roughly 200 BC and 650 AD, are so large that they are fully visible only from the air, making systematic exploration difficult and slow. Despite many theories about their purpose — from astronomical markers to ritual pathways — their meaning and full extent have remained unclear.
To accelerate discovery, a team led by Professor Masato Sakai of Yamagata University in Japan, in collaboration with researchers from IBM, turned to artificial intelligence. They trained AI systems to scan and analyze aerial and satellite imagery of the Nazca Desert, searching vast regions for patterns that are too subtle or faint for human observers to spot easily. In just six months, this technology identified 303 previously unknown figurative geoglyphs — nearly doubling the number known to researchers.
The use of AI drastically reduced the time and effort required to survey hundreds of square kilometres of desert. What would have taken archaeologists years of ground-based work was achieved in a matter of months, demonstrating how machine learning can process high-resolution images quickly and thoroughly. According to Sakai, this efficiency allows researchers to focus field efforts on validating potential finds rather than manually combing the landscape.
While the new discoveries bring scholars closer to understanding the full scope of the Nazca Lines, questions about their purpose — whether ritualistic, astronomical, or otherwise symbolic — remain unresolved. Nonetheless, the success of AI in this project highlights the growing role of advanced technology in archaeology, helping uncover lost human heritage across the globe as more historical data becomes digitised and accessible.
