The Algorithmically Positioned Self: Rethinking Dialogicality in Human-AI Interaction

AI platforms are creating new conditions for meeting the voices with which people come to understand themselves. One easy way out of this problem is to conclude that the use of AI constitutes a new type of Other, or a new I-position, in dialogical self theory (DST). This paper takes issue with such a conclusion and proposes that AI produces a unique semiotic context for communication - utterances that are responsive and appear to be other-generated, but that are not generated by an embodiment or historically positioned subject such as human dialogue assumes in DST. Dialogicality thus does not lie in the AI; it arises from the interplay between the individual person, the AI-produced utterance, and the person's repertoire of culturally mediated positions. Based on DST, semiotic-cultural psychology, and work on algorithmic curation and AI companions, the paper offers a human-system configuration approach to AI interaction wherein technological factors (ranking systems, interfaces, goals, and generated language) are taken up into meaning transformation by human beings. It makes the distinction between human agency and system effect, highlights the difference between high agency and low agency using three analytical examples, incorporates the culture into the theory, and explains how it contributes to understanding of the algorithmic, datafied, and distributed self. It ends with limited, testable hypotheses and a revised discussion of the clinical and educational significance of the theory. Its ultimate objective is to understand what the emergence of AI teaches us about I-position, Other, and dialogue according to DST.