An Irreducible Agency Invariant: Technical Briefing

This 6-page technical briefing distills the core computational structure of the Irreducible Agency Invariant (IAI), providing the essential definitions, equations, and causal criteria required to test for agency in recurrent systems. It is designed as a companion to the full IAI paper and offers a compact entry point for AI safety researchers, neuroscientists, computational cognitive scientists, and philosophers of mind seeking a rigorous, non-metaphysical criterion for identifying internally authored transitions in artificial or biological agents. For correspondence please email: mikeferketic73@gmail.com