Criminal Responsibility for Artificial Intelligence-Related Crimes under International Criminal Law: A Comparative Study of the Legal Systems of Iran and Iraq

Abstract

Artificial intelligence increasingly shapes military operations, security governance, surveillance, data analysis, cyber activities, and automated decision-making, creating new challenges for the attribution of criminal responsibility. This article examines criminal responsibility arising from artificial intelligence-related crimes under international criminal law and comparatively evaluates the legal systems of Iran and Iraq. Using a doctrinal, analytical, and comparative methodology, the study investigates whether existing principles governing individual criminal responsibility, causation, mens rea, complicity, indirect perpetration, aiding and abetting, command responsibility, and corporate liability can adequately address crimes committed through, facilitated by, or resulting from artificial intelligence systems. The analysis shows that artificial intelligence should not presently be treated as an independent bearer of criminal responsibility because current systems lack moral agency, legal consciousness, and meaningful responsiveness to punishment. Instead, responsibility should remain centered on human and organizational actors who design, authorize, deploy, supervise, control, or knowingly benefit from high-risk systems. Existing international criminal-law doctrines can address many intentional and knowing uses of artificial intelligence, particularly where the technology functions as an instrument of genocide, crimes against humanity, war crimes, or other serious offences. However, significant accountability gaps arise when harmful outcomes result from opaque algorithms, autonomous learning, distributed decision-making, negligent development, or fragmented institutional control. The comparative analysis indicates that Iranian and Iraqi law contain general rules concerning intention, negligence, causation, perpetration, complicity, omission, and legal-person liability that may partially govern artificial intelligence-related harm. Nevertheless, neither system provides a comprehensive framework for high-risk artificial intelligence, autonomous military systems, algorithmic evidence, or responsibility throughout the technological life cycle. The article proposes a multi-layered model allocating liability among intentional users, operators, developers, manufacturers, corporate managers, military commanders, and civilian superiors according to their knowledge, control, contribution, risk creation, and failure to prevent harm. It concludes that legislative reform should preserve personal culpability while introducing clear duties of human oversight, testing, auditability, documentation, cybersecurity, and evidence retention.