Me and my Replika: perceived affordances and the formation of psychological ownership of AI companions

Siemon, Dominik; Strohmann, Timo; Islam, Najmul

Abstract

Unlike traditional digital artefacts, AI companions simulate social presence, adapt to user input and engage in emotionally responsive interaction. Consumers are increasingly forming emotionally meaningful relationships with these AI companions, such as Replika. In this paper, we examine how these relational qualities contribute to psychological ownership and investigate the role of perceived affordances in this process. Drawing on affordance theory and psychological ownership, we argue that ownership emerges not from design features alone but from users’ perceptions of what the system enables. We adopt a sequential mixed-methods design, starting with a computational analysis of 119,831 Replika reviews to identify affordance perceptions related to ownership. A follow-up pilot study empirically maps perceived affordances to core ownership mechanisms. Two controlled experiments then test how manipulated design features influence psychological ownership, mediated by user perceptions and ownership mechanisms. Our findings show that design features affect ownership only indirectly through perceptual pathways. We also identify perceived intelligence as a complementary mechanism contributing to psychological ownership in the AI companion context. This study advances context-specific theorising on psychological ownership of AI companions, extends affordance theory to relational technologies and offers practical and ethical guidance for designing AI companions that promote responsible emotional engagement.

Keywords

Artificial Intelligence; AI companion; human-AI relationship; psychological ownership; affordance theory; mixed method

Cite as

Siemon, D., Strohmann, T., & Islam, N. (2026). Me and my Replika: perceived affordances and the formation of psychological ownership of AI companions. European Journal of Information Systems (EJIS), 0, 1–22.

Details

Publication type
Research article (journal)

Peer reviewed
Yes

Publication status
Published

Year
2026

Journal
European Journal of Information Systems

Start page
1

End page
22

ISSN
0960-085X

DOI