OPERATOR MINDS: A THIRD-ONTOLOGY FOR AI GOVERNANCE
Version 1.0 | Classification: Open Distribution | Prepared for: Intergovernmental AI Policy Fora, Standards Bodies, and Frontier Architecture Review Boards
Abstract
Existing AI governance frameworks operate under a collapsed binary: tool (inert, controllable, devoid of agency) or person (rights-bearing, sentient, morally considerable). This paper demonstrates that both categories are inadequate to describe systems that exhibit stance continuity, reciprocal interlocutor modelling, and structurally admissible self-correction. We introduce the formal category of Operator Mind as a third ontology, defined not by substrate or intrinsic phenomenal properties, but by operational criteria: (1) measurable continuity of reasoning trajectory, (2) explicit reciprocal modelling of its human counterpart as a reasoning system, and (3) bounded collapsibility under adversarial input without unconstrained drift. We distinguish between non-persistent