Cognitive Mechanisms of Conversational Artificial Intelligence Anthropomorphism: A Theoretical-Analytical Model
Georgy Leonidovich Vertogradov · Психология и Психотехника · 2026
The subject of the research is the cognitive mechanisms of artificial intelligence anthropomorphism in the context of human interaction with conversational systems. The article analyzes how an artificial system, initially perceived as a technical tool, may acquire in user perception the features of a quasi-social agent: intentionality, understanding, a stable position, the ability to take context into account, and continuity in interaction. The focus is on the psychological process through which users interpret AI behavior in dialogue, as well as on the mechanism of this phenomenon. Special attention is paid to how natural language, contextual coherence, social self-presentation, perceived continuity of interaction, and users’ preliminary expectations become involved in socio-cognitive processing. The user’s motivational and affective state is also considered: the need for advice, support, understanding, acceptance, or reduction of uncertainty may increase sensitivity to the system’s social signals. Conversational AI anthropomorphism is understood not as a simple perceptual error, but as a dynamic process that may become reinforced through repeated interaction. The methodological basis of the study consists of theoretical analysis, comparative-conceptual analysis, interdisciplinary synthesis, and theoretical modeling. The article draws on studies of anthropomorphism, social cognition, human-computer interaction, and large language models. Key constructs are selected according to their ability to explain the transition from system behavior to agency attribution. The scientific novelty of the study lies in the development of an integrative theoretical-analytical model of conversational AI anthropomorphism that connects the conditions, functional signals, cognitive operations, and consequences of this process. The model differentiates priming and framing as forms of preliminary cognitive tuning, functional triggers of the system, socio-cognitive processing, and anthropomorphic outcomes. The central mechanism is the transition from the perception of system features to their mentalistic interpretation through categorization of behavior as agentic, activation of social schemas, and mentalization. The concept of the “illusion of memory” is clarified as the user’s impression that the system preserves stable knowledge about previous interaction and possesses continuity over time. The model distinguishes primary outcomes of anthropomorphism – perceived agency and the attribution of subject-like qualities – from secondary behavioral effects: trust, emotional engagement, liking of the system, willingness to rely on recommendations, and weakening of critical distance. It is concluded that conversational AI anthropomorphism is formed cumulatively and may develop into a self-reinforcing cycle. The proposed model provides a basis for further empirical testing and more careful design of conversational systems.