the distance between “dog” and “fan,” indicating that “dog” and “cat” represent more similar concepts. Emergence Emergence refers to a phenomenon in which a complex system displays new properties or capabilities that arise from the interactions of simple components—properties or capabilities that were not anticipated during the design phase. In generative AI, the term describes situations in which large language models, after being trained on massive datasets with billions or trillions of parameters, (appear to) develop abilities that were never explicitly programmed by developers. Examples often cited include the ability to provide logically structured answers to certain questions or to transform text into a variety of styles. These emergent behaviors can manifest not only in positive ways but also in negative ones, such as generating misinformation or facilitating hacking-related tasks. Whether such capabilities truly represent new abilities is still under debate, but the discussion highlights the fact that there remain gaps in our understanding of how generative AI works and how its behavior can be predicted. Anthropomorphism Excessive anthropomorphism of generative AI can obscure the technology’s fundamental nature and limitations, creating various risks. It is therefore important to clarify that generative Generative AI Guide for Civil Society