Abstract
With the advancement of virtual and augmented reality technologies, as well as distributed computing platforms, there has emerged a need for the creation of digital entities capable of operating independently and interacting with other participants in a virtual environment. These entities are known as autonomous AI agents – software components capable of making decisions, adapting to changes, and collaborating within the framework of multi-agent systems (MAS). In the context of the metaverse, the role of such agents goes beyond purely service-oriented functions. They can act as NPCs (non-playable characters), assistants, trading intermediaries, educational systems, or moderators, as well as operate as independent entities with long-term objectives. The purpose of this article is to examine the architecture, operational mechanisms, and prospects for the use of autonomous AI agents in multi-user virtual worlds, as well as to present visual diagrams illustrating key concepts.