Artificial intelligence (AI) systems, like all other technologies, are not mere instruments but processes to be developed. However, unlike all other technologies, their autonomous functioning allows them to have a stronger active contribution in the interaction processes with the users, raising new challenges for the law. Since AI systems may be involved in different ways in the commission of a crime, in the future, it is likely that existing offences will have to be adapted or new AI-based crimes created. One of the main issues that the autonomy of AI systems poses to criminal law is who is to be held criminally responsible in case of harmful events caused by their emergent behaviors. In these cases, a responsibility gap could follow. The artificial agent cannot be held directly responsible and the human agent, who has no full control over the system’s autonomous functioning, cannot always be criminally reprovable for not exercising the duty to act required to him. According to the new regulatory framework recently proposed by the European Institutions (the so-called Artificial Intelligence Act), the paper aims to describe the possible and future criminal policies that will allow avoiding a responsibility gap.
Artificial Intelligence and criminal law: the myth of “control” in a data-driven society
Beatrice Panattoni
2021-01-01
Abstract
Artificial intelligence (AI) systems, like all other technologies, are not mere instruments but processes to be developed. However, unlike all other technologies, their autonomous functioning allows them to have a stronger active contribution in the interaction processes with the users, raising new challenges for the law. Since AI systems may be involved in different ways in the commission of a crime, in the future, it is likely that existing offences will have to be adapted or new AI-based crimes created. One of the main issues that the autonomy of AI systems poses to criminal law is who is to be held criminally responsible in case of harmful events caused by their emergent behaviors. In these cases, a responsibility gap could follow. The artificial agent cannot be held directly responsible and the human agent, who has no full control over the system’s autonomous functioning, cannot always be criminally reprovable for not exercising the duty to act required to him. According to the new regulatory framework recently proposed by the European Institutions (the so-called Artificial Intelligence Act), the paper aims to describe the possible and future criminal policies that will allow avoiding a responsibility gap.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
