This paper is a structural diagnostic analysis of coercive regimes under AI mediation. It operates strictly downstream of Informational Ontology (revision 5i) and Witnessing Under Vulnerability. All definitions, exclusions, and regime commitments from those works are assumed as authoritative. The analysis is non-normative. It does not argue that any regime is justified or unjustified. The central question is whether AI-mediated pipelines preserve the attributional structure required for coercive regimes to remain regimes rather than dissolving into logistical throughput systems. The paper examines how AI mediation thins coercive regimes by fragmenting responsibility across pipelines such that no single participant can be said to have imposed the coercive outcome, even when the outcome is formally lawful and procedurally compliant. This thinning occurs without error, malfunction, or rule violation—it appears as efficiency, consistency, and procedural success while the witness disappears. What replaces law is not tyranny or automation, but logistics: constraint without termination, force without exposure, and deprivation without a bearer of responsibility.
Michael Semprevivo (Tue,) studied this question.