L'intelligenza artificiale apre una nuova era negli studi cognitivi dei primati selvatici

An AI system designed to study the cognitive abilities of wild capuchin monkeys has been introduced, marking a significant advancement in primate research. Developed by scientists at Emory University and the Georgia Institute of Technology, the technology combines facial recognition and real-time touchscreen testing to enable automated cognitive assessments in natural habitats. This innovation was detailed in a proof-of-concept study published in The American Journal of Primatology. The system, called CapuchinAI, features a compact, battery-powered computing unit integrated into a field research platform. It identifies approaching monkeys using facial recognition, presents personalized learning tasks via a touchscreen interface, and rewards successful responses with food. Field trials conducted in the Taboga Forest Reserve of Costa Rica demonstrated the system’s effectiveness, achieving 97% accuracy in identifying individual capuchins after being trained on still images and video footage. The monkeys quickly adapted to the touchscreen-reward setup, showing rapid habituation and learning capabilities. Capuchin monkeys are renowned for their intelligence, intricate social structures, and curious nature, making them excellent subjects for studying the evolution of cognition. The new AI tool enables researchers to systematically assess and track cognitive functions in these animals within their natural environments, a challenge previously hindered by the difficulty of maintaining controlled conditions in the wild. Researchers emphasize that this approach offers a scalable solution for gathering consistent cognitive data across large populations. The project draws inspiration from the pioneering work of Frans de Waal, a leading figure in primate cognition research who served as director of Emory’s Living Links Center for the Advanced Study of Ape and Human Evolution. De Waal’s contributions included both laboratory experiments and observational studies that highlighted the complexity of primate behavior and social dynamics. His influence is evident in the current effort, which seeks to combine technological precision with the nuanced understanding of individual primate experiences. Federico Sánchez Vargas, the lead author of the study and an Emory Ph.D. candidate in anthropology, notes that the AI method complements traditional field observations by allowing for automated cognitive testing. “We can now quantify findings and gain deeper insights into the minds behind the personalities,” he explains. “Studying individuals in their natural environments, where life experiences vary widely, helps us understand how environmental factors shape their cognitive abilities.” Training the facial recognition component required extensive manual input from undergraduate students, who meticulously annotated faces in field videos and assigned correct identities to each monkey. The individuals involved in the study were named Pele, Pedro, Pio, and Pesto. This level of detail underscores the importance of human oversight in refining the AI’s ability to recognize and interact with specific subjects. The research team includes Jacob Abernethy, an associate professor of computer science at Georgia Tech, and Sai Rakshith Potluri, a former graduate research assistant at Georgia Tech who is now employed as a software engineer at ExtraHop in Seattle. Their collaboration highlights the interdisciplinary nature of the project, blending expertise in artificial intelligence with field research methodologies. The open-source publication accompanying the study includes comprehensive documentation of the code used to develop the system, alongside instructions for constructing a low-cost, low-tech field research platform. The authors aim to encourage other researchers to adopt and expand upon this methodology, potentially applying it to study cognitive traits in diverse species of wild primates across varied ecosystems. Benítez’s work bridges disciplines such as anthropology, psychology, and evolutionary biology, focusing on themes like cooperation and social behavior. Her ongoing research continues to explore how cognitive capacities emerge and evolve in response to environmental pressures, offering valuable insights into the broader mechanisms of human and non-human cognition.