AI Insight
This theoretical article argues that when emerging adults use conversational AI for identity formation, the technology occupies a unique "synthetic position" within the self that differs from human interlocutors by being protean, frictionless, simulated in opposition, and authored elsewhere. The authors propose that habitual reliance on AI for self-formative decisions can lead to "authorial suspension," where individuals have many internal voices but lack authentic self-authorship of their identity. Using Dialogical Self Theory, they develop a conceptual model predicting that the structural role AI plays in identity formation matters more than frequency of use, which may explain why population-level studies show weak effects.
Why it matters
This framework has implications for understanding how AI tools may fundamentally alter identity development in young adults, not just accelerate existing processes. The authors suggest this could inform therapeutic approaches and educational policies around AI use during critical developmental periods.
Understand the Science
Generative artificial intelligence (AI) has been widely treated as an instrumental aid to thought, a faster route to information, drafting, and articulate self-description. This article problematizes that framing and argues that, when enlisted for the intimate work of deciding who to become, conversational artificial intelligence alters the structure of emerging adulthood, not only its content. Working within Dialogical Self Theory, which conceives the self as a society of relatively autonomous I-positions integrated by a promoter position, we propose that a large language model in self-formative dialogue occupies the self as a synthetic position: an interlocutor that is protean (able to voice any position on demand), frictionless (built to accommodate rather than resist), simulated in its counter-positioning (offering the form of opposition without a lived stake), and authored elsewhere (generating positions instead of returning the user’s own positions). We argue that these four properties are jointly and reliably present to an extent that is only ever partial in a human other, a combination the existing typology does not anticipate. When habitually enlisted, the synthetic position tends to absorb the integrative authorship the theory assigns to the meta-position and the promoter position, a process we term synthetic promotion, producing a condition we name authorial suspension: a self densely voiced yet un-authored. We render this account as a conceptual model, distinguish authorial suspension from the neighboring constructs with which it might be confused, and translate the mechanism into formal, falsifiable propositions. Chief among them is the prediction that the outcome turns on the structural location the synthetic position is permitted to occupy, not on the volume of use, so that population-level studies indexing use alone should record a washed-out average effect even where the mechanism operates. Taking a recent Chinese graduating cohort as an illustrative case, we specify the measures an empirical test would require and draw implications for constructivist therapy and education.