AI Claims Maya Discovery, But Experts Point to Prior Findings

Miscommunication in Archaeological Discoveries

A recent claim by Google researcher Pietro Miotti suggested the discovery of a new Maya site in Mexico, powered by AI technology. However, this assertion was met with swift criticism as archaeologists pointed out that the site in question had already been documented in an academic journal more than a year ago.

The AI Investigation

Miotti's journey into archaeology began during a trip to Peru, where he pondered how historical explorers might utilize modern technology. His solution involved using lidar, a laser mapping tool that helps create detailed topographical images. Engaging a team of 48 AI agents, he analyzed existing NASA lidar data for potential undocumented sites, shifting his focus to the Maya region when he found public data on Peru lacking.

The AI agents flagged numerous locations for further investigation, including a significant village which Miotti believed hadn't been officially registered. Though he stated this project was conducted as a personal initiative, he contacted Mexico's National Institute of Anthropology and History (INAH) to verify the status of his findings. Using AI translation tools, Miotti communicated in Spanish with INAH, which initially informed him that they had no record of a site at that location.

Experts Weigh In on the Controversy

The incident ignited discussion on platforms like Bluesky, where archaeologists emphasized the necessity of comparing new findings with existing research. Sarah Parcak, an archaeology professor, remarked that a lack of registration in a government database does not equate to an undiscovered site. This sentiment was echoed by colleagues who criticized the potential spread of misinformation by enthusiastic but untrained individuals.

Miotti clarified that as a non-archaeologist, his understanding of archaeological research processes was limited. He acknowledged the confusion surrounding site registrations and the intricate task of verifying known sites, asserting the importance of expert involvement in the interpretation of AI findings. He quickly updated his original blog to credit the work published by professional archaeologists.

AI’s Role in Future Archaeological Work

Despite the criticism, Miotti advocates for the benefits AI can bring to experts by processing large datasets. He noted a positive response from various archaeologists and PhD students eager to explore how AI can streamline their workflow. Yet, Elizabeth Graham, an expert in Mesoamerican archaeology, emphasized that the primary issue lies in the funding necessary to explore many sites already on the radar of archaeologists, many of which are known but remain unexcavated.

Additionally, Ivan Šprajc, an archaeology professor, expressed skepticism regarding the significance of the AI detectives' work, asserting that the archaeological community has been employing lidar and remote sensing tools long before this incident. He highlighted that merely identifying features does not enrich the existing body of knowledge and cautioned against the influx of new discoveries from amateurs.

Implications for the Future

The episode raises broader implications about the intersection of AI and scientific research. While AI tools boast impressive capabilities, their application by non-experts can lead to oversights, misunderstandings, or misrepresentations of established science. This situation underscores the essential role of experienced professionals in interpreting findings to ensure accurate reporting of discoveries.

As the debate continues, archaeologists urge responsibility in the use of emerging technologies, advocating for collaboration between experienced researchers and AI developers. The promise of AI lies in its ability to enhance existing methodologies, but without proper context and expertise, its potential could remain unfulfilled.