Standalone jurisprudential research work
Artificial Agency Without Personhood
Toward a Jurisprudence of Autonomous AI
Version 1.0 · 10 September 2026 · Preprint, not peer reviewed
Purpose
Making legally relevant architecture visible when autonomous systems select consequential means.
Artificial Agency Without Personhood examines whether established legal doctrines can coherently attribute conduct and allocate responsibility when autonomous objective-directed systems pursue objectives, select intermediate means and produce legally consequential effects.
Its central position is that law can recognise Functional Agency without converting an artificial system into a rights-bearing or duty-bearing legal person. The analysis remains grounded in human and institutional responsibility across the system lifecycle.
Analytical contribution
A coordinating legal framework for conduct produced through autonomous means selection.
The paper treats objectives and constraints as a single legally relevant architecture and develops Under-Constrained Success to describe cases in which effective objective attainment exposes inadequate constraints on permissible means.
It distinguishes Authorised Reach, Required Technical Reach and Provisioned Technical Reach, and develops coordinating concepts concerning system lineage, acquired capability, Risk-Class Foreseeability, Evidentiary Responsibility and the Distributed Causal Enterprise.
These concepts are offered as provisional analytical tools. They do not create artificial personhood, presume universal liability or collapse distinct legal systems into a universal rule.
Relationship to Machine Bylaws
Independent in legal purpose, structurally related in the problem it illuminates.
Machine Bylaws addresses the governance and control conditions surrounding consequential non-human agency. Artificial Agency Without Personhood examines the adjacent legal question: how conduct, risk and responsibility may be attributed when the consequential means are selected autonomously.
The jurisprudential paper is a complete standalone work. It supports legal understanding of the wider agentic-governance problem, but it is not part of AAGA and does not announce a continuing jurisprudence programme or publication series.
Publication record
J-01 · Zenodo · 10 September 2026
Artificial Agency Without Personhood
Toward a Jurisprudence of Autonomous AI
Recommended citation: Bjorn Bjornsvik (2026). Artificial Agency Without Personhood: Toward a Jurisprudence of Autonomous AI (Version 1.0) [Preprint]. Zenodo.
Defined concepts