Abstract

Artificial intelligence has revolutionized policing by transforming massive amounts of data into valuable, actionable insights. Its most prominent applications include predictive policing, facial recognition, digital forensics, and decision-support systems In this context, this study aims to demonstrate the role of artificial intelligence and digital authorities in crime prevention and control, while focusing on monitoring the risks of losing control over these systems and defining the dimensions of criminal liability arising from them. The risks of human-machine interaction—such as algorithmic bias and the opacity of complex decisions—constitute a major obstacle to the effectiveness of artificial intelligence systems. This imposes ethical and human rights challenges that necessitate the establishment of strict legal frameworks to protect international peace and security. These robots and systems remain merely artificial tools owned by humans to expand their capabilities and achieve their goals; thus, they should neither be anthropomorphized nor held legally or ethically liable for their actions. However, artificial intelligence faces the risk of "brittleness" and the inability to adapt to the real world due to the "distributional shift" problem, where systems fail drastically when encountering environmental or systemic changes that differ from the data they were trained on. To achieve these objectives and explore the nature and applications of this technology, the study adopts a descriptive-analytical approach to deconstruct the mechanisms of the smart system. The study concludes with the necessity of filling the legislative vacuum regarding the limits of human control over smart technologies to ensure the accuracy of attributing criminal liability procedures.

Keywords

Artificial Intelligence; Criminal Liability; Policing; Crime Prevention; Algorithmic Bias; Distributional Shift; Descriptive-Analytical Approach.