From Disclosure to Self-Referential Opacity: Six Dimensions of Strain in Current AI Governance

TL;DR

This paper examines how AI governance opacity evolves with increasing system capabilities, applying a six-dimension political theory framework to analyze current governance arrangements and their limitations.

Contribution

It introduces a six-dimension framework from political theory to analyze AI governance opacity and applies it to real-world arrangements, highlighting the limits of transparency remedies.

Findings

Proprietary secrecy decreases with capability asymmetry but leads to gaming or embedded evaluation.

Legitimacy and non-domination are more consistently strained than corrigibility and resilience.

Institutional design quality influences the responsiveness of certain governance dimensions.

Abstract

Governance opacity over AI systems shifts in kind as capability asymmetry grows, and the strongest forms defeat the disclosure-based remedies governance ordinarily relies on. This paper applies a six-dimension framework from political theory (legitimacy, accountability, corrigibility, non-domination, subsidiarity, institutional resilience) to six AI governance arrangements already in operation, ordered by increasing capability asymmetry between system and overseer. Proprietary secrecy yields to disclosure at the low end, but at the high end the governed system either games its own evaluation or sits inside the governance process, and transparency remedies lose traction. Legitimacy and non-domination strain more consistently across the sample than corrigibility and resilience, which respond more readily to institutional design quality. The sample cannot separate institutional design maturity from capability asymmetry, and the patterns are offered as hypotheses for multi-rater validation.