A recent project utilizing autonomous AI agents to scan satellite and laser data for historical ruins has ignited a spirited debate across the scientific community. Google researcher Pietro Miotti recently claimed that a team of 48 AI agents he deployed had discovered a previously unmapped Maya site in Mexico. However, specialists in Mesoamerican archeology quickly pointed out that the supposed discovery was entirely known, having been formally documented in an academic journal more than a year prior.
Miotti initiated the personal side project without any Google compute resources or official backing, driven by a curiosity about how historical explorers might leverage modern technology. He utilized lidar, a laser-based mapping method capable of generating three-dimensional environmental models, to scour NASA data. After shifting his focus from Peru to the Maya region due to data availability, his AI agents flagged several promising anomalies, including what appeared to be an unregistered village.
Key takeaways
- Google researcher Pietro Miotti used 48 AI agents and NASA lidar data to search for unmapped Maya sites in Mexico.
- The AI agents flagged a location that Miotti claimed was an undocumented village.
- Professional archeologists revealed the site had already been discovered and published in an academic journal last year.
- Experts warn that amateur AI projects risk spreading misinformation and overlooking existing academic literature.
Seeking initial validation, Miotti contacted Mexico’s National Institute of Anthropology and History to ask if the coordinates corresponded to a registered site. Because the government agency had no official record in its database, Miotti publicized the finding on his blog, framing it as an example of artificial intelligence revolutionizing scientific discovery. He later admitted that he did not realize an unregistered site in a government database could still be fully documented in peer-reviewed literature.
Professional archeologists took to social media to voice their frustrations. Sarah Parcak, an archeology professor at the University of Alabama at Birmingham, emphasized the vital role of peer review and warned against assuming that a lack of governmental registration equals an unmapped discovery. Other scholars noted that local communities are often already well aware of features in their surrounding areas, and that the term discovery is frequently misused when applied to modern remote sensing.
Elizabeth Graham, a professor of Mesoamerican archaeology at University College London, pointed out that the fundamental bottleneck in archeology is a lack of funding for excavation and fieldwork rather than a shortage of spotted mounds. Ivan Šprajc, a coauthor of the paper that originally documented the site, expressed concern that a surge of amateur AI claims could overwhelm working researchers. Šprajc also noted that publicizing precise coordinates carries the inherent risk of attracting looters to sensitive historical locations.
Despite the backlash, Miotti acknowledged the limitations exposed by the exercise, noting that specialized tools must evolve to better compare findings against figures locked inside academic papers rather than simple databases. While experts maintain that rigorous domain expertise remains indispensable, the incident highlights a broader tension as technologists experiment with specialist fields.
Comments