§ 11. Attribution of Algorithmically Co-caused Action

This page accompanies the German section § 11. Zurechnung algorithmisch mitverursachten Handelns. It treats attribution as a social operation rather than as a simple reconstruction of technical causality. Users, organizations, developers, data chains, models, and oversight bodies form distributed constellations in which responsibility may be assigned, displaced, or obscured.

The section therefore rejects both easy personification of AI and the opposite claim that systems are merely neutral tools. Models are not moral subjects; when they are said to decide, discriminate, or hallucinate, this is a practical shorthand for a wider human-machine arrangement. The core problem is the institutional organization of responsibility under many-hands conditions, responsibility gaps, and auditability.

In the architecture of Chapter I, attribution links closure and access with algorithmic associations. The question becomes: who may deploy a system, who must understand it, who can contest its outputs, and who has the authority to recognize harm as harm?