The Changing Nature of Human-AI Relations: A Scoping Review on Terminology and Evolvement in the Scientific Literature

Recent years have brought immense progress in the development of AI technology. This broadened its application fields but also led to a surge of interest in many research domains and increasing significance of human-AI relations for the development of AI technology. This rapid growth and evolvement is reflected by the establishment of a great variety of terms, potentially leading to what is known as jingle and jangle fallacies. With our scoping review of the terminology used in scientific literature to describe human-AI relations and its evolvement over time (with 803 records screened, 658 finally included), we capture the variety and development of human-AI terminology in accordance with the shift from <i>interaction</i> to <i>collaboration</i> between humans and AI. We aim to raise awareness of these developments spanning over different research communities and provide a solid basis for future researchers and practitioners conducting complementary, cross-domain research. Our review comprises terminological, bibliometric and thematic analyses, e.g., reporting on the historical development of terms and term composition patterns, but also identifying key authors and publications, geographic distribution of relevant research, and elaborating on term conception and usage, and co-occurrences throughout the literature.