Discerning Undeclared AI-Generated Data in OpenStreetMap

This study assesses the differentiation between human-generated and AI-generated road data in collaborative mapping. A metric, focusing on node density variability, is introduced. Applied to Microsoft AI and OSM volunteer-generated road data, AI data shows uniformity, while human-generated data displays significant variability. Recognizing distinctive patterns allows identification and differentiation of AI and human road data, potentially addressing undeclared data origins in OSM.

Veja a apresentação no congresso State of the Map US