Pietro Miotti reports that 48 AI agents examined roughly 1,200 square kilometres of NASA lidar data, with less than 72 hours from his initial idea to an INAH reply. In a November 11 correction, he acknowledged that a 2025 paper had already mapped the location. It was absent from the institute’s register, but the agents missed the literature and rediscovered a known site. The experiment suggests potential for terrain screening, not successful novelty verification; interpretations of the site’s age and purpose remain speculative.
Back to Tech
AI Maya discovery claim corrected after agents missed an existing paper
Unregistered does not mean undiscovered; parallel analysis still needs literature checks and expert verification.