AI & Alignment

Vulnerability Laundering in AI-Mediated Regimes

In production

A diagnosis of how AI-mediated regimes can preserve accountability in form while systematically preventing consequence from returning to commitment-capable loci in a trajectory-binding way.

Back to corpus

Function in corpus

Names a specific structural failure mode in AI-mediated governance where auditability, attribution, and traceability remain visible while the exposure needed for endogenous correction is laundered away.

Details

The paper distinguishes accountability in form from accountability in function. It shows how aggregation, routing, fragmentation, and interface translation can leave adverse consequence fully legible while structurally eliminating return paths by which that consequence could become trajectory-binding at commitment-capable loci. The result is a regime that appears increasingly accountable while losing the structural basis of correction. • Connected papers: Salience Misalignment; Vulnerability as a Structural Condition; Witnessing Under Vulnerability; Legitimacy as a Finite Structural Resource; Working as Designed; Interruption Authority Under Scale

Availability

This paper is listed for orientation and dependency tracking. No public PDF or Zenodo record is linked yet.