Autonomous Weapons Based on Artificial Intelligence and Human Control: Challenges and Solutions in International Humanitarian Law

Articles in Press

Document Type : Research Paper

Authors

1 Ph.D. in Islamic Jurisprudence and Law, Assistant Professor, Department of Jurisprudence and Law, Imam Khomeini and Islamic Revolution Research Institute, Tehran, Iran

2 Master's degree in International Law, Payam Noor Law School, Tehran, IranØ› Graduated from the Department of Islamic Jurisprudence and Law, Faculty of Theology, Al-Zahra University, Tehran, Iran. 1-2 Graduated Masters Degrees from the Department of International Law, Faculty of Law Payame Noor University,

10.22099/jls.2026.55440.5456

Abstract

Artificial intelligence-based autonomous weapons have fundamentally transformed decision-making processes on the battlefield, thereby generating novel challenges for International Humanitarian Law (IHL). Machine learning technology facilitates the automation of critical decisions; however, the continuous learning of these systems through environmental feedback renders the prediction of their behavior and the analysis of their decision rationale difficult and at times impossible, challenging the feasibility of adhering to international rules and wartime requirements. Through a descriptive-analytical method, this article examines the technical capacities and legal implications of autonomous weapons, aiming to propose solutions for reconciling these two ostensibly contradictory domains. It demonstrates that the deployment of such weapons necessitates a redefinition of human control to both ensure compliance with international legal frameworks and exploit the technology for mitigating wartime casualties. The findings indicate that the appropriate design and responsible deployment of machine learning systems can enable the better implementation of cardinal principles of IHL, such as military necessity and the distinction between military targets and civilians.

Keywords

Main Subjects

Journal of Legal Studies

Articles in Press, Accepted Manuscript
Available Online from 10 February 2026

Files

Share

How to cite

Statistics