Temporal authorship as a moral right: bounded inquiry in persistent AI delegation

Existing AI ethics frameworks—focused on fairness, privacy, and accountability—do not currently provide an adequate conceptual vocabulary for the moral harm of temporal domination: the institutional foreclosure of a person’s future self by treating past algorithmic classifications as permanently authoritative. This article articulates and defends a moral right to temporal authorship, defined as the capacity to escape how one has been classified and author one’s own future. We ground this right in a diachronic conception of personhood, distinguish it from existing concepts of autonomy and epistemic injustice, and derive from it four institutional duties—intelligibility, justification, contestability, and temporal exit—that serve as necessary conditions for the morally permissible delegation of persistent, life-consequence classifications to AI systems. We demonstrate the novelty and plausibility of this framework through a comparative analysis with existing approaches and through case studies of real-world AI systems, including the Dutch childcare benefits scandal (SyRI) and recidivism risk assessment in the U.S. criminal justice system. The article concludes by outlining the governance implications of temporal authorship, including the need for temporal audits and expiration triggers for algorithmic classifications.