Temporal authorship and the moral illegitimacy of pre-authorized AI: a new ethical architecture for delegated inquiry

Artificial intelligence systems increasingly make consequential decisions that bind individuals’ future selves to judgments made about their past selves. This paper introduces a clarificatory and integrative normative framework to address a significant ethical threat arising from this practice: temporal domination. We argue for Temporal Authorship as a distinct diachronic standing claim: the right of persons to author their own lives across time, free from the indefinite binding of their identity to historical patterns by algorithmic systems. This paper puts forward a normative account that offers a powerful framework for governing delegated inquiry. We show that existing AI ethics and governance frameworks, while valuable, systematically under-specify this temporal dimension by treating authorization as a one-time event. We derive four non-compensatory conditions for protecting Temporal Authorship—epistemic transparency, normative justification, institutional contestability, and temporal reversibility—which together constitute Bounded Inquiry, a new normative architecture for the democratic governance of delegated inquiry. This framework advances a defensible governance condition: delegation without continuous renewal is presumptively illegitimate. By shifting the focus from regulating system behavior to governing system authority, Bounded Inquiry provides a new, legitimacy-preserving foundation for the ethics of AI.