Advanced artificial intelligence is increasingly incorporated into interdependent work in ways that blur the familiar distinction between organizational tool and social collaborator. Existing scholarship explains human–AI teaming, algorithmic control, anthropomorphism, social presence, trust, and professional-role identity, yet these literatures remain fragmented when workers begin to interpret AI as a relationally consequential participant in work. This Original Conceptual Theory Article develops the proposed Theory of Artificial-Coworker Relations (TACR) to explain how AI may become construed along a tool–artificial-coworker boundary and how that construal may connect social-relational and identity processes. TACR proposes that AI affordances, work-role conditions, and human or team characteristics shape relational object construal; that social presence and repeated interaction may support workplace emotional attachment; and that work-role overlap, comparison, substitution, or complementarity may make AI interaction a site of proposed identity projection. Boundary management concerning role definition, decision rights, responsibility, reliance, disclosure, override, and escalation is theorized to condition these processes. The theory distinguishes anthropomorphism from social presence, trust from attachment, professional-role identity from identity projection, and technological agency from legitimate authority. Its propositions are conditional and falsifiable rather than claims of validated causality. The framework is bounded by uneven cross-context evidence, especially the transfer of attachment findings from social-chatbot settings to work. Validation requires longitudinal workplace research, discriminant construct development, experimental manipulation of relational cues and authority boundaries, and multilevel tests capable of challenging the theory.
