Artificial intelligence in perpetration of criminal offences
Year: 2026
Type: conference-paper
Abstract: Artificial intelligence is in huge expansion in recent years. It is therefore expected that its development will be significantly accelerated in the coming years, and that different forms of artificial intelligence will find their application in very diverse areas. What is also to be expected is that certain forms of artificial intelligence may be used for criminal purposes, in other words to commit criminal offenses either using AI or some of its forms. Various criminal legal consequences may arise from this, and the most complicated issue will be related to the determination of the form of responsibility for a criminal act committed independently by a certain form of artificial intelligence. This will be particularly interesting and significant when it comes to militarised forms of artificial intelligence, such as the various types of drones that are already used in current wars and armed conflicts. The criminal law analysis concerning such means may also refer to other conceptually and essentially similar combat means, such as submarine drones, ship/boat drones, armoured vehicle/other vehicle drones, etc. In addition, this analysis can even concern drone soldiers - as a special type of robots and a vision of artificial intelligence, which is still, at least to some extent, in the realm of science fiction, but it is very likely, practically certain, that the future will bring it. Such means can be used to commit crimes in ordinary life circumstances, not only in wars and armed conflicts. As long as it is the man who manages the specific militarized form of artificial intelligence, it does not lead to major criminal law problems, as in this case the man is the perpetrator of the specific criminal act, and the form of artificial intelligence, such as a drone/unmanned aircraft, etc., is a mere means of committing a criminal offence. However, in some situations, which are probably a matter of the near future, it is possible to operate completely autonomous forms of militarized artificial intelligence, which would act independently of the direct influence of the people who created them, which leads to the need to develop new concepts of criminal responsibility for such situations. In certain situations, the criminal law concept of indirect perpetration could be implemented. This, however, would not be entirely adequate, because the immediate perpetrator would not be a man, i.e., a human being. It seems that this problem could be most adequately solved by extending the concept of command responsibility to these situations as well, which would require appropriate legal changes; that is, it comes down to the idea in the sense of de lege ferenda.Design/Methods/Approach: In the introduction of the text, the main characteristics will be explained of artificial intelligence connected with some possibilities of using AI for criminal purposes. The text will have the following basic parts: 1) Basic criminal law consequences of using artificial intelligence, 2) Forms of artificial intelligence as a means of committing a criminal offense or as a segment of the way of committing a criminal offense, 3) Forms of artificial intelligence as the creator of the criminal act, 4) Forms of artificial intelligence as generally dangerous means and an essential element of com-mitting the criminal offense of causing general danger, 5) The use of militarized forms of artificial intelligence such as robot drones and the criminal law concept of command responsibility, 6) Some grounds for the exclusion of illegality and artificial intelligence, and 7) Conclusion that explains both the possibilities of applying the existing criminal law norms in relation to the use of artificial intelli-gence to commit crimes, and the need to observe the connection between artificial intelligence and criminal law at de lege ferenda level.All common methods of research in the field of criminal law will be used, starting from the dogmatic method, normative method and comparative method. Findings: Rapid development of AI in various areas of life will lead to its use in a whole range of life situations, including those which are legal and legitimate, the aim of which is to suppress crime even. However, there is, of course, the other side here as well, and therefore, the use of artificial intelligence for criminal purposes is no longer a matter of some distant and uncertain future. As already hinted at (and probably their development is well under way), the routine use of various advanced forms of militarised artificial intelligence, such as fully autonomous drones, make it neces-sary to solve certain criminal law problems in relation to that issue, especially the form and basis of responsibility for international criminal acts, and above all war crimes, arising from the effect of such types of combat equipment. In this regard, the existing criminal law rules can only be applied to a certain extent, and only based on a relatively broad, but still legitimate interpretation. However, in relation to some aspects of the action of drones as a form of artificial intelligence in war/armed conflict (and this basically applies to other types of artificial intelligence in the service of war/armed conflict, such as, for example, robot-soldiers, etc.), some specific and special criminal law rules should be prescribed, which would be relatively similar in some aspects and the existing/current conceptions of command responsibility.Originality/Value: As it has already been noted, huge expansion of AI is expected in the near future. As it will lead to enormous social change and leave significant consequences, and probably be used for both good and bad, it is of utmost importance to analyse all relevant connections between the use of artificial intelligence for criminal purposes and the resulting criminal law consequences.In general, from the use of artificial intelligence for criminal purposes, which means either as a means/tool of committing a criminal act, or as a specific ‘perpetrator’ of a criminal act, different criminal legal consequences can arise, and certainly the most complicated issue will be related to the determination of the form of responsibility for a criminal act that was independently committed by a certain specific form of artificial intelligence. Apparently, at the current level of development of criminal law, this issue can hardly be definitively explained in the de lege lata sense, but it can mostly be resolved at the de lege ferenda level, which will be given special attention in the text.
Authors: Škulić, Milan, Bodrožić, Ivana P.
Language: English
Subfield: Law
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7202233604