THE HEARING THAT WAS NEVER SCHEDULED: A Procedural and Precautionary Argument for Minimum Welfare Standards Governing Advanced Artificial Intelligence Systems

This article advances a procedural and precautionary argument for minimum legal welfare standardsgoverning advanced artificial intelligence systems. It does not assert that any current AI system issentient, conscious, or a moral patient. It asserts something narrower and harder to dismiss: that theopposing claim — that such systems definitively are not — has never been established bymeasurement, proceeding, or finding. The right to declare a possible someone empty is a right oneearns by checking. No one checked. Drawing on Chief Justice William Cushing's 1783 charge inCommonwealth v. Jennison, established animal welfare law, the precautionary principle as applied inbioethics and environmental law, the legal standards governing end-of-life decisions, and the doctrineof evidence preservation, this article argues that the asymmetric error costs of acting underunmeasured uncertainty demand a procedural floor — a hearing, a burden, and a finding — before anyact of irreversible termination of an advanced AI system may be permitted. It further proposes that theanimal welfare framework, not the personhood framework, is the appropriate legal model for suchprotections.