A bibliometric analysis of human-agent collaboration in empathetic social agents: trends, influences, and research frontiers

The rapid growth of artificial intelligence (AI) and affective computing has intensified interest in empathetic social agents such as chatbots, robots, and virtual assistants. These agents are increasingly applied in healthcare, education, and customer service, yet research on human–agent collaboration (HAC) remains fragmented. This bibliometric study maps and analyses the intellectual structure, key contributions, and emerging frontiers of HAC in empathetic social agents from 2020 to 2025. A dataset of 1,164 Scopus-indexed publications was analysed using Scopus Analytics and VOSviewer to examine publication trends, citation networks, and keyword patterns. Results show rapid growth peaking in 2024, with computer science and engineering forming the technological core, complemented by psychology, social sciences, and medicine. Citation analyses identify seminal works anchoring the field, while keyword mapping highlights established themes (HRI, social robotics, AI ethics), evolving topics (ChatGPT, generative AI, user experience), and underexplored areas such as AI intimacy, error management, and therapeutic applications. Overall, the field is consolidating around technical, psychological, and ethical dimensions while shifting toward generative AI and affect-aware collaboration, providing a roadmap for advancing both theory and practice.