Researchers have uncovered 303 new geoglyphs in Peru’s Nazca Desert, thanks to the use of artificial intelligence (AI). This discovery, made by a team led by Japanese archaeologist Masato Sakai from Yamagata University, doubles the known number of these ancient etchings and sheds light on their potential functions, deepening our understanding of the mysterious Nazca Lines.

The geoglyphs, massive drawings carved into the desert floor over 2,000 years ago, are spread across a vast 400-square-kilometre area. They include linear figures, geometric shapes, and humanoid forms. Cutting-edge AI technology was used to analyse high-resolution satellite and aerial images, allowing researchers to detect smaller, more difficult-to-spot geoglyphs that had eluded previous expeditions.

The latest discovery includes 303 relief-type geoglyphs, which differ from the larger, more famous line-type figures. While many of the line-type geoglyphs depict wild animals such as condors and monkeys, a significant number of the newly discovered figures include human-like figures and scenes that may represent human sacrifice. The researchers also noted that these relief-type geoglyphs are often found near ancient pathways, suggesting they may have been created to be seen by travellers, possibly serving a ceremonial or spiritual purpose.

This new understanding comes after decades of research into the Nazca Lines, which were first brought to international attention in the 20th century. Unfortunately, the site has also been subject to damage. In 2014, Greenpeace activists caused outrage when they staged a protest near the famous hummingbird geoglyph, leaving behind footprints that inflicted lasting damage on the fragile desert floor. The incident served as a stark reminder of how vulnerable the site is to human activity.

AI’s role in archaeology, particularly in preserving such delicate cultural heritage, is becoming increasingly significant. In the case of the Nazca Desert, AI has allowed researchers to accelerate the discovery process, which would have otherwise taken years using traditional methods.