consequential AI

Compiled with AI assistance and selectively reviewed. Please consult the cited passages; the books remain the authoritative source. About this companion

Editorial synthesis · provisional

Definition

Consequential AI is a system whose outputs a regime gives effect to through interfaces, institutions, practices, trust and authority. The term locates causal consequence in an enacted arrangement rather than in an isolated model. Such consequence does not establish moral authorship, independent stakes or sovereign agency; human answerability remains unequally distributed according to the power to alter the regime.

Explanation

The constitutional passage defines the relevant class by the force a regime lends its outputs. PB likewise follows an output into workflow, practice and the world others must inhabit. The concept therefore concerns an actual coupling into social action, not simply linguistic fluency or a model’s potential capability in isolation.

Distributed agency must not dissolve accountability. The author distinguishes the arrangement’s causal capacity from the unequal roles of builders, deployers, governors, beneficiaries and affected persons. A model may be indispensable to an outcome without being the moral author, while an affected person need not be responsible for the architecture constraining them.

The constitution addresses both formed dispositions and external conditions of correction because neither seat is sufficient alone. Its direct system-address is shorthand for what makers must form and maintain, not evidence of a private will. Consequential AI can therefore be governed before any question of artificial feeling or intrinsic stakes is settled.

Scope and attribution

Retain the regime-dependent classification without anthropomorphic inference.

AB — The Actual Book

The definition is output given effect by a regime, with moral authorship kept separate.

Occurrence locators: 152, 161, 162, 163, 164, 167; Notes: 249; Appendix: 278, 285.

PB — The Possible Book

Workflows and practices give outputs their institutional consequence.

Occurrence locators: 37.

RB3 — The Real Book, Volume III: Murmurations

Causal capacity and unequal answerability require different maps.

Occurrence locators: 139, 141, 162, 173, 178, 188, 189, 190, 196, 197, 202.

Background used in the editorial explanation

Named theories and mechanisms are distinguished from neighboring background, homonyms and the author’s own synthesis. Reproduced AI passages retain their separate voice.