Professor Dr. Ammar Abbas Al-Husseini<br />Dean of the College of Law<br /><br />In light of the widespread, diverse, and rapidly evolving use of artificial intelligence technologies and applications, and in the absence of clear legislation to address the crimes and damages that may result from such technologies, it has become urgent and necessary to discuss the crisis of criminal law provisions in confronting AI-related crimes. This raises questions about the adequacy of current criminal legislation to address and encompass the unique features of AI technologies. This issue manifests in several key points, including:<br /><br />What is meant by "criminal law provision"? And what is the principle of legality in criminal law?<br /><br />Are traditional penal provisions sufficient to criminalize acts involving AI?<br /><br />Can existing legislation related to cybercrime in various countries be utilized in this context?<br /><br />How can these legal provisions address crimes committed by AI entities themselves—especially in cases where AI systems evolve autonomously without human intervention?<br /><br />The term criminal law provision refers to the structure in which a criminal rule is formulated, or the set of terms that define offenses and their penalties. It represents the legal vessel through which the criminal legislator enforces their will upon the subjects governed by it. Criminal law provisions are characterized as imperative rules of public law. Accordingly, the relationship they govern arises between the state and individuals who violate these provisions. Being imperative, neither individuals nor authorities have the right to alter them or agree to bypass them.<br /><br />It is important to note that criminal law provisions are not limited to penal codes alone but also appear in various other legislations that include punitive measures, such as laws against prostitution, drug laws, juvenile care laws, and cybercrime laws. Generally, the criminal law provision is the outcome through which the principle of legality in criminal law is embodied, wherever it may appear.<br /><br />As for the principle of legality—nullum crimen, nulla poena sine lege (no crime and no punishment without a law)—this principle represents the legal element of a crime. It means that the illegality of an act stems from its conformity with a legal text that defines it as a crime. Hence, the only source of criminalization and punishment is the criminal law provision itself. Without it, there can be no talk of a "crime," regardless of the extent of the harm caused or the social condemnation it may provoke. Customs, morals, or social views do not in themselves create a crime. Criminalization and punishment are among the most serious legislative issues as they directly affect individuals’ rights and freedoms. Therefore, the principle of legality is now the cornerstone of the entire criminal justice system, and as such, analogy is strictly prohibited in criminalization.<br /><br />The expanding scope of crimes committed through or by AI raises a significant challenge: efforts to criminalize some of these behaviors may clash with the principle of legality. In practice, this leads to several possible scenarios:<br /><br />Some AI-related crimes may fall under the general scope of traditional criminal laws, such as incitement or blackmail committed through AI systems. These can potentially be prosecuted using general provisions that do not consider the means by which the crime was committed, in line with the notion that criminal law focuses on the act, not the tool.<br /><br />Some offenses may be covered under cybercrime laws, as AI systems are software-based, and such laws may already address crimes committed through or against software. However, very few countries—including Iraq—currently have such legislation in place, even as of the time of writing.<br /><br />Existing general principles of traditional criminal law, such as liability, error, or corporate responsibility, may be employed to address certain acts committed by AI systems.<br /><br />Despite these avenues, the coming sections will explore the various assumptions surrounding AI crimes, keeping in mind the legal and practical challenges involved. One major issue is the frequent violation of privacy rights by AI, which creates a new category of offenses requiring criminalization. A more pressing concern is determining who bears responsibility for crimes committed autonomously by AI systems, especially as many of these systems now have the capacity to evolve and act independently of human input. This raises a new legal problem that current laws struggle to address. A striking example is Microsoft’s AI chatbot Tay, which, in 2016, sent thousands of racist and illegal messages within eight hours via what was then known as Twitter (now X). Another case involves the chatbot Eliza, which reportedly drove an individual to commit suicide.<br /><br />In countries like Iraq, where no dedicated cybercrime legislation exists and where the penal code has not been amended to include AI or cyber-related offenses, the judiciary has no authority to invent new crimes—owing to the principle of legality. This makes the enactment of new legislation or the amendment of existing laws to address AI crimes an urgent matter.<br /><br />Hence, the justification for legislative intervention becomes clear—especially since AI is now an unavoidable reality due to its expanding applications and the increasing number of crimes associated with it, many of which are beyond human control. This necessitates a robust criminal legislative response.<br /><br />Al-Mustaqbal University – The Top University in Iraq