Answers are grounded in this paper and cite their sources.
The increasing sophistication of generative artificial intelligence has transformed human–computer interaction from predominantly instrumental exchanges into communication experiences that can resemble interpersonal dialogue. Contemporary conversational AI systems respond to natural language, maintain contextual coherence, simulate conversational empathy, and adapt their responses to user preferences. These capabilities create psychological conditions in which users may perceive an artificial system not merely as a technological instrument but as a digital interlocutor possessing apparent social qualities. This conceptual paper examines five interconnected dimensions of human–AI communication: anthropomorphism, perceived social presence, trust formation, interpersonal interpretation, and psychological expectations. Drawing on social cognition, human–computer interaction, and research on trust in automation, the paper investigates how conversational characteristics influence users' perceptions, emotional responses, self-disclosure, and willingness to rely on AI-generated information. A central argument is that the psychological experience of communication with AI should be distinguished from the technological and ontological properties of the system itself. Users may experience understanding, support, companionship, or interpersonal responsiveness without these experiences establishing the existence of human-like consciousness, emotional reciprocity, or independent social intentions within the AI. The paper introduces the Perceived Digital Interlocutor Framework (PDIF), a conceptual model explaining how technological affordances, individual psychological characteristics, social attribution, and expectation formation interact to shape human–AI relationships and behavior. The framework emphasizes the distinction between productive social engagement and potentially maladaptive forms of overtrust, emotional dependence, and excessive attribution of human characteristics. It also identifies opportunities for enhancing reflective thinking, psychological accessibility, and calibrated trust. The proposed framework provides a theoretical foundation for future experimental, longitudinal, and psychometric investigations of human–AI communication, particularly in educational, professional, supportive, and psychologically sensitive contexts
No takes yet. Share an insight, caveat, or question.
Roman Shaban (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: