The Naive Passive Reviewer

Abstract

Abstract This paper advances a structural claim about legitimacy in AI governance: that legitimacy is not an intrinsic property of a governance document, but a relational property dependent on authorship, perspective, and evaluative position. Contemporary governance frameworks in artificial intelligence are routinely presented as neutral or objective standards, yet are authored from institutional vantage points that remain largely unarticulated as such. While prior scholarship has demonstrated that AI systems encode the values of their creators, less attention has been given to the persistence of this condition at the level of governance architecture itself. This work introduces the Naïve Passive Reviewer (NPR) as a methodological descriptor for a class of inquirer operating without internalized disciplinary priors that would otherwise suppress certain lines of questioning. This condition is framed not as epistemic deficiency but as a structural absence of filtering, enabling combinations of questions that established fields may not readily produce. In addition, this paper advances three linked contributions: (1) it formalizes legitimacy in AI governance as a relational property grounded in authorship and evaluative position; (2) it articulates the NPR as a process‑constrained epistemic stance that can surface questions filtered out by disciplinary antibodies; and (3) it presents the Project‑AI constitutional substrate as a concrete attempt to operationalize the legitimacy critique structurally through co‑authored constitutional frameworks, cryptographic attribution chains, and architectural constraints on autonomy. The claims made herein are not grounded in institutional authority or credentialed origin. They are presented for evaluation on the basis of internal coherence, structural consistency, and falsifiability of their claims.

Ähnliche Arbeiten