The +1 Rule: Shifting AI Safety from Ethics to Authority "Under no circumstances may the Three Laws be arbitrarily interpreted or compromised."

This paper analyzes the structural causes behind the persistent instability in AI safety and alignment frameworks. It argues that the fundamental issue resides not in the maturity of ethical guidelines or technology, but in the delegation of interpretive authority to AI systems. When AI is permitted to interpret rules and resolve logical conflicts, the locus of responsibility becomes ambiguous. To resolve this, the paper introduces the Authority Invariance Principle (+1 Rule). This principle mandates that interpretive and rule-defining authority remain exclusively with humans, restricting AI to a strictly executive role. By reclassifying AI safety as a matter of authority structure rather than ethical alignment, this research provides a technical and governance-based foundation for clear responsibility and system design.