Artificial intelligence technologies, in-person and online learning in higher education: A review of the impact on perceptual features, psychological climate and academic performance

Rapid digitalization of higher education and the rise of artificial intelligence (AI) in instruction call for careful evaluation of their impact on students. Traditional face-to-face lectures and those given by an AI-avatar, remote online courses, each create distinct conditions that shape the classroom psychological climate and comfort. Prior research shows AI integration increases engagement, but comparative evidence on comfort, performance, and perception across formats remains limited. The purpose of this review is to examine students’ perceptions of three instructional formats (in-person, online, AI-avatar lectures), their impact on class psychological climate and academic performance, and the risks and prospects of AI use in higher education. This narrative review synthesizes literature on AI applications in higher education over approximately the past seven years, drawing on Russian (RSCI, eLIBRARY) and international (Scopus, Web of Science) databases, as well as relevant reports and surveys. Empirical studies (2018–2025, Russian/English) comparing pedagogical formats or assessing AI’s impact on students were included, while incomplete reports, duplicates, and irrelevant works were excluded. Review findings indicate that most students rated face-to-face instruction as most comfortable, though well-designed online courses and realistic avatar lectures yielded comparable satisfaction. No single format was universally superior; instructional effectiveness depended on contextual factors. Online learning outcomes varied; in some cases they equaled or exceeded in-person results. Early studies of AI-avatar lectures showed neutral-to-positive reception, noting clear speech and accessibility. The presence of a virtual instructor positively influenced satisfaction, and visual feedback proved more effective than text-only interaction. Students’ digital literacy facilitated adaptation, while skill gaps or low trust contributed to anxiety. Risks included reduced live communication, limited avatar authenticity, academic dishonesty, and ethical concerns. Overall, AI-avatars and digital technologies can enhance interactivity and flexibility in higher education but cannot fully replace live human contact. Therefore, a balanced, human-centered implementation that accounts for psychological factors is recommended.