TOWARDS AN INTERNATIONAL CRIMINAL-LAW FRAMEWORK FOR ARTIFICIAL INTELLIGENCE IN DEFENCE AND SECURITY: A PROPOSAL FOR A NEW CONVENTION

Authors

  • Mirsulton Abbosov Independent Researcher, Tashkent State University of Law, Uzbekistan Author

Keywords:

Artificial intelligence, criminal law, international criminal law, autonomous weapons, human control, criminal responsibility, actus reus, mens rea, artificial intelligence agency, international convention.

Abstract

The rapid development of artificial intelligence (AI) has created new challenges for international security, international humanitarian law, and criminal law. The use of increasingly autonomous AI systems in military and defense activities creates a serious legal problem: when an AI-enabled system causes unlawful death, injury, or other serious consequences, who should bear criminal responsibility? Existing international instruments provide important rules concerning the conduct of hostilities and the protection of human rights, but they do not establish a sufficiently specific criminal-law framework for highly autonomous AI-enabled weapons and systems. This article examines this legal gap and proposes a new international Convention on the Regulation of Artificial Intelligence Technologies and Their Use in Defense and Security. The proposed Convention was developed within a doctoral research project and consists of 17 articles. Its central principles include a prohibition of fully autonomous weapon systems, mandatory and effective human control over permitted AI-enabled weapons, international monitoring, algorithmic audit, explain ability, emergency shutdown mechanisms, and direct linkage between actus reus, mens rea, human conduct and individual responsibility. The article argues that the main criminal-law innovation of the proposed Convention is not the recognition of artificial intelligence as an independent criminal subject, but the establishment of a clear human-centered responsibility model for crimes committed with the involvement of autonomous technologies. The article also discusses the proposed International Artificial Intelligence Agency as an institutional mechanism for monitoring compliance, conducting inspections and investigating serious incidents. The study concludes that a specialized international instrument could complement existing AI governance and international humanitarian law by addressing the criminal-law dimension of autonomous military technologies.

Downloads

Published

2026-05-31

Issue

Section

Articles