Brent Carr

Recent debates about AI “guardrails” in mental health have framed a problem as one of detecting delusion. This paper argues that such framing misunderstands both psychiatry and large language models. In clinical practice, delusion is not simply a false belief but a structure of conviction that stabilizes meaning under conditions of uncertainty. Large language models, by contrast, generate coherence rather than evaluate truth. When these systems interact with users seeking orientation, they do not merely fail or succeed at detection—they participate in the conditions under which belief forms. The ethical risk is therefore not misclassification but premature coherence: the subtle reinforcement of interpretations before they can be examined. Drawing on phenomenology and clinical psychiatry, this paper proposes that safety in AI systems should be understood as the protection of provisionality—the preservation of the human capacity to remain uncertain long enough to think.

My argument: P1: If my personhood is caused by Truth then, Truth is a person. P2: My personhood is caused by Truth. C: Therefore, Truth is a person. The argument hinges on the Causal Principle: A cause is, at least, as powerful as its effect. I can get 1 from 10, but I cannot get 10 from 1; therefore, the causal principle is vindicated. Truth is more than a person as you and I are, He is the archetype of personhood by and in which all other personhoods move and have their being. Unlike Descartes who was unable to stop a slide into ever more skepticism, the Archimedean point of Truth - though, I am not directly going into math and logic here I believe them to function likewise as bulwarks against skepticism – does just that. The tsunami of skepticism has been unable to sweep up this hill. We can indeed know with surety, which is absolute certainty, that Truth exists and by extension God, who is Truth, and the rest of what Descartes builds back as sure i.e., the external, sensory world.