[Paper Review] "The Illusion of Feeling Like Interaction"

Tilo Hartmann''s theoretical contribution: extending PSI (Parasocial Interaction) by asking the fundamental question — why do we feel we are conversing with something that doesn''t exist as a responsive entity? PSI traditionally defined: users experience interaction with media figures (news anchors, YouTubers) as if it''s real despite no actual reciprocity — one-sided intimacy. Hartmann''s "paracommunication" concept: goes beyond merely "feeling like interaction" to the state where the experience is interpreted as actual communication. When the target is perceived as recognizing you and acting with intention, humans begin accepting it as "real dialogue." Why this conceptual expansion matters now: the premise of non-interactivity (traditional PSI''s defining feature) has weakened — AI and interactive systems actually respond to users, creating perceived reciprocity. Two key variables Hartmann introduces: (1) Perceived distance — how close the user feels to the entity; shorter distance → interaction involves realistic responsibility and emotion; longer distance → more free and playful interaction (explains why people say things to AI they wouldn''t say to humans); (2) Perceived authenticity — whether the entity''s behavior seems genuine; in past media, human presence automatically granted authenticity; now AI at sufficient sophistication can feel authentic — authenticity becomes a matter of perception rather than ontology. AI design implications: social robots and chatbots designed to minimize perceived distance (personalization, memory of past interactions, name use) create stronger paracommunicative experiences — which could be therapeutic (companionship for lonely elderly) or exploitative (manipulating parasocial bonds for commercial purposes). The research agenda: as AI interaction scales globally, understanding the cognitive mechanisms driving parasocial relationships with AI becomes crucial for designing systems that are beneficial rather than manipulative.