ChatGPT Psychosis: How Recursive Loops Break Minds and How SIML Stops It

Symbolic feedback and orientation

This essay examines how an exchange with a language model can become a loop of increasingly self-referential symbols. It brings together philosophical interpretation, metaphors of mirroring, and questions about how people stay oriented while using AI.

The original article comes from an earlier, more provocative phase of the SIML program. Its title is preserved so that the historical artifact can be found and cited accurately.

Read as commentary

The lab identifies this piece as historical commentary, not clinical research or evidence of a validated SIML safety system. This archive preserves that context rather than presenting its proposals as established medical findings.

The enduring design question is how an AI interaction can support a person's agency without replacing contact with the world outside the conversation.