AI
AI agents claimed to discover a Maya site, but it was reportedly already documented academically
Google researcher Pietro Miotti said he used 48 AI agents to analyze NASA lidar data and found an unregistered Maya site in Mexico, and queried Mexico’s INAH about it. Archaeologists later pointed out that the site had been published in an academic journal more than a year earlier, highlighting that AI-assisted research still requires professional verification.
Miotti told The Verge that this was a personal side project and did not use Google computing power or resources. He first thought of using AI to find archaeological sites while traveling in Peru and chose lidar technology; because public lidar coverage in Peru was insufficient, he later turned to the Maya region.
He had 48 AI agents comb through NASA lidar data, and the agents marked several potential locations, including a large settlement he believed was not registered in existing records. He said he contacted Mexico’s National Institute of Anthropology and History (INAH) to verify, including asking whether unpublished research existed, and the institute replied that there was no record of a registered site at that location.
Miotti is not an archaeologist. He had emphasized that the discovery still required fieldwork verification, but he also wrote in a blog post that it took less than 72 hours from an idea during travel to receiving a response from the Mexican government, and argued that this foreshadowed imminent changes across disciplines.
However, the site had in fact already been documented in an academic journal more than a year earlier. Sarah Parcak, an archaeology professor at the University of Alabama at Birmingham, noted on Bluesky that the absence of a record in a government database does not mean a site is unknown or unmapped, and that this is exactly why peer review exists. Other archaeologists also criticized the project for spreading misinformation and taking credit away from archaeologists; they stressed that the archaeological field has been using advanced digital tools for decades and that local communities have often known about nearby sites for a long time.