AI Alignment and the Minimal Alignment Operator
Abstract
This work reframes AI alignment as a problem of human coherence and presents a minimal alignment operator based on invariant preservation, closure cycles, and the engine–shape boundary. It defines alignment as the stability of a system’s internal structure under expansion and offers a formal predicate capturing the conditions under which an intelligent system remains self‑stabilizing.
Fields of Study & Subjects
Ethics and Social Impacts of AI
Psychiatry
Mental Health
Neuroscience
Cybernetics and Technology in Society
Publication details
- Journal / source
- OSF Preprints (OSF Preprints)
- Publication date
- 31 March 2026
- Language
- EN
- Document type
- Preprint