Artificial intelligence (AI)-driven avatars within virtual reality (VR) environments have become pivotal in advancing human-computer interaction (HCI) by offering immersive, interactive, and personalized experiences. This systematic literature review examines 20 peer-reviewed studies published between 2020 and 2025, sourced from leading academic databases including IEEE Xplore, ACM Digital Library, Scopus, SpringerLink, and ScienceDirect. The review focuses on the state-of-the-art AI techniques and technologies employed in developing intelligent avatars, their impact on user interaction, and the challenges faced in real-world applications. Employing thematic analysis, the study identifies key trends such as the use of deep learning, natural language processing, and embodied conversational agents, alongside issues of privacy, ethical considerations, and computational constraints. Results reveal that while AI-driven avatars significantly enhance VR experiences across domains such as education, healthcare, and entertainment, challenges around scalability and user acceptance persist. This review concludes by outlining future research directions to improve avatar adaptability, ethical frameworks, and multimodal interaction capabilities. These insights aim to guide researchers and practitioners toward developing more effective and user-centered AI avatars in virtual environments.
Kohistani et al. (Wed,) studied this question.