When adaptation stops being a strategy: a hypothesis on the internalization of cognitive processes in prolonged interactions with conversational artificial intelligence

Abstract

This article proposes a theoretical reflection on the possibility that prolonged interactions with conversational artificial intelligence systems promote, in the user, a gradual process of cognitive and communicational adaptation. While the available literature extensively documents the adaptation of the system to the user, the hypothesis defended here shifts the focus toward the reverse direction, proposing a reorganization model structured in three progressive layers: an epistemic filter, responsible for the defensive pre-qualification of the user’s own subjective judgments; an aesthetic-expressive filter, which conditions symbolic and figurative production to a prior stage of critical validation; and a possible extra-contextual generalization, in which this self-regulation pattern extends to the subject’s cognitive processing beyond the context of interaction with artificial intelligence systems. Grounded in Communication Accommodation Theory (Giles, 1973) and the extended mind thesis (Clark; Chalmers, 1998), and in dialogue with emerging research on bidirectional human-AI adaptation, the proposition does not present empirical results, but offers an initial conceptual vocabulary for a phenomenon still dispersed in the literature, indicating directions for future empirical investigation through longitudinal analysis of linguistic production and comparative experimental designs.