UDTR — Radar mondial des droits de la Nature

From Human Offenders to Artificial Agents: Toward a Criminological Framework for Autonomous AI-Enabled Crime

Identification

Clé publique

doi:10.62225/2583049x.2026.6.5.6914

Auteurs

Brandon Presley · Muhammad Arslan

Année déclarée

2026

Langue déclarée

en

Type source

article

Première détection

2026-11-09T19:04:47.713286+00:00

Dernière observation des métadonnées

2026-11-09T19:04:47.713286+00:00

Dernière vérification HTTP

Non effectuée

Institutions déclarées par la source

Institute of Southern Punjab · Purdue Global

Pays d’affiliation déclarés

PK · US

Résumé documentaire — langue originale

“Generative and agentic artificial intelligence (AI) is developing at a rapid pace, changing the relationship between technology and crime. However, the growing autonomy of systems that can plan, adapt, use tools and take goal-directed actions, challenges this traditional human-centric perspective and the dominant conceptualization of AI in existing research as a tool that enhances human offending. This narrative literature review explores the shift from human led AI assisted crime to human-AI collaborative and potentially autonomous artificial agent offending. We synthesize literature from criminology, cybersecurity, artificial intelligence, criminal law, and technology governance to evaluate the relevance of established criminological perspectives, including rational choice theory, routine activity theory, social learning theory, deterrence, and situational crime prevention. Particular attention is paid to emerging challenges such as artificial agency, criminal intent, responsibility attribution, multi-agent interactions and machine-oriented deterrence. The review provides a criminological framework for conceptualizing AI-enabled crime on a continuum ranging from traditional human offending to AI-assisted crime, human-directed agentic crime, human-AI collaboration, semi-autonomous AI-enabled crime, and, in theory, highly autonomous artificial-agent crime. We propose six dimensions to analyze these new forms of offending: human involvement, AI autonomy, goal formation, decision-making capacity, foreseeability and responsibility attribution. The review argues for a criminology that moves beyond a purely human-centred conception of crime, but also avoids prematurely attributing consciousness or human-equivalent culpability to artificial intelligence systems. Future empirical research should take into account human-AI co-offending, multi-agent interactions, computational guardianship, forensic attribution and the applicability of criminological theories in autonomous digital environments. This framework provides a basis for comprehending crime in increasingly complex sociotechnical human-machine systems.

Identifiants

OPENALEX

W7214173096

DOI

10.62225/2583049x.2026.6.5.6914

Repères documentaires

Famille documentaire

Doctrine

Types déclarés par les sources

article

Les mentions d’entités ne constituent pas une preuve de reconnaissance juridique. La juridiction étudiée est distincte du pays d’affiliation des auteurs.

Sources

OpenAlex · Document 1 · Document 2

• DOI · https://doi.org/10.62225/2583049x.2026.6.5.6914
• OpenAlex · https://openalex.org/W7214173096

Statut documentaire

Pertinence à examiner

Bibliothèque

Cette identité n’est pas encore intégrée à la Bibliothèque juridique.