Lions as Bone Accumulators? Exploring Multi‐Predator Contributions to the Olduvai Carnivore Site (OCS) (Tanzania) Through AI and Metric Analyses

Lions ( Panthera leo ) are apex predators with a well‐documented influence on ecological dynamics, yet their potential role as bone‐accumulating agents remains poorly understood and often debated. Previous taphonomic studies have largely attributed bone accumulations in African savannah ecosystems to other carnivores, such as spotted hyenas ( Crocuta crocuta ) and leopards ( Panthera pardus ), due to their known prey‐transporting and bone‐modifying behaviors. However, some research has suggested that under certain ecological conditions, lions may also contribute to the formation of bone assemblages, thereby challenging long‐standing assumptions about faunal assemblage formation processes. That is the case with Olduvai Carnivore Site (OCS), a modern wildebeest bone assemblage from Olduvai Gorge attributed to nomadic lions. Although the lion accumulation hypothesis proposed for OCS was supported by coherent and internally consistent lines of taphonomic and ecological evidence, several variables in the assemblage suggest that its formation may have been more complex than initially recognized. In this study, we conducted a re‐study of this assemblage to reassess the accumulating agents involved, applying artificial intelligence (AI) techniques—specifically deep learning algorithms trained on high‐resolution images—to analyze and classify tooth marks across the collection. By comparing these AI‐derived patterns with reference datasets from known carnivore‐modified assemblages, we were able to evaluate the likelihood of lion agency and identify potential post‐depositional modifications by other agents. We combined this approach with metric analysis of tooth pits. Felids and hyenids generate divergent tooth pit sizes on dense bone from long bone midshafts. Combining both approaches, we showed—contrary to previous analyses—that hyenid was the predominant agency in bone modification. The results demonstrate the potential of AI‐driven taphonomic analysis to refine interpretations of bone surface modifications (BSM) and carnivore behavior, with significant implications for interpreting Plio‐Pleistocene archaeological and paleontological sites.