CRIMINAL LIABILITY AND ARTIFICIAL INTELLIGENCE: THEORETICAL AND APPLIED ASPECTS

The humanity is now at the threshold of a new era when a widening use of artificial intelligence (AI) will start a new industrial revolution. Its use inevitably leads to the problem of ethical choice, it gives rise to new legal issues that require urgent actions. The authors analyze the criminal law assessment of the actions of AI. Primarily, the still open issue of liability for the actions of AI that is capable of self-learning and makes a decision to act / not to act, which is qualified as a crime. As a result, there is a necessity to form a system of criminal law measures of counteracting crimes committed with the use of AI. It is shown that the application of AI could lead to four scenarios requiring criminal law regulation. It is stressed that there is a need for a clear, strict and effective definition of the ethical boundaries in the design, development, production, use and modification of AI. The authors argue that it should be recognized as a source of high risk. They specifically state that although the Criminal Code of the Russian Federation contains norms that determine liability for cybercrimes, it does not eliminate the possibility of prosecution for infringements committed with the use of AI under the general norms of punishment for various crimes. The authors also consider it possible to establish a system to standardize and certify the activities of designing AI and putting it into operation. Meanwhile, an autonomous AI that is capable of self-learning is considerably different from other phenomena and objects, and the situation with the liability of AI which independently decides to undertake an action qualified as a crime is much more complicated. The authors analyze the resolution of the European Parliament on the possibility of granting AI legal status and discuss its key principles and meaning. They pay special attention to the issue of recognizing AI as a legal personality. It is suggested that a legal fiction should be used as a technique, when a special legal personality of AI can be perceived as an unusual legal situation that is different from reality. It is believed that such a solution can eliminate a number of existing legal limitations which prevent active involvement of AI into the legal space.