Criminal Liability of Artificial Intelligence Crime in Indonesia: Challenges and Opportunities

Mahrus Ali1, Muhammad Oscar Dharma Putra Mulya2 & Wahyu Priyanka Nata Permana3

Abstract
This study explores the challenges and opportunities for Artificial Intelligence to be criminally liable for offenses in the Indonesian Criminal Legal System. This study employed doctrinal legal research focusing on the legislation and various secondary legal sources of the relevant topics. It is argued that three (3) models of criminal responsibility for this matter, which depend on the role of the AI in the commission of a crime can be applied; (1) perpetration via another, (2) natural probable consequence, and (3) direct liability model. It is also not impossible for an A.I. to be pinned with criminal responsibility, as is the case with natural persons and corporations. The factual requirement, the actus reus, can be satisfied with no issue in the event that an A.I. operates its body parts to commit an act, including assault towards a person. The mental element or the mens rea, can be fulfilled by treating the systems of A.I. the same as the human brain; both A.I. and the human brain process and analyse information and data obtained through receptors. However, the challenge that may be raised is that no matter how logically a rule is set up, it holds no water if it is not approved or accepted by society. Unfortunately, it cannot be determined yet whether these models of punishment intended to be imposed upon A.I. will suit the needs of the society or not, as again, Indonesia is not in the realm of Society 5.0 yet, and thus the theory regarding the imposition of criminal responsibility, along with the types of sanction which may be pressed upon it, is still hypothetical.

Keywords: Artificial intelligence, Crime, Criminal responsibility, Society 5.0