Percorsi per la protezione delle persone fisiche dalle inferenze create dall’intelligenza artificiale: right to explanation, right to reasonable inferences e artificial intelligence act.
Abstract: Paths for the protection of individuals from inferences created by artificial intelligence: right to explanation, right to reasonable inferences and artificial intelligence act – Artificial intelligence systems using machine learning techniques are able to discover probabilistic correlations between sets of data they are trained with. They are, therefore, able to make predictions and assessments, inferring new data, called, in fact, inferences. Inferences can be defined as information about an identified or identifiable natural person, created through deduction or reasoning, rather than simple observation or collection from the data subject. However, they are probable but inevitably uncertain information, for which neither the data used as training-set nor the procedure followed by the system is known. In addition, inferences depend on the input data: they can only be reliable if the data on which they are based are reliable. In spite of the many risks, the use of inferences is growing exponentially and forms the basis of decisions that affect the economic and social lives of individuals, without any European law directly addressing them. The purpose of this paper is to analyse the European legislation, already in force or still being drafted, concerning the protection of personal data and artificial intelligence, as well as the solutions put forward by the academic world, in order to ascertain whether they can be applied with respect to inferences and whether they provide an adequate protection.
Sommario: 1. Introduzione. – 2. L’intelligenza artificiale e le inferenze. – 3. I rischi determinati da un uso distorto dell’intelligenza artificiale.– 4. Le inferenze e la legge sulla protezione dei dati personali. – 5. Le inferenze e le altre normative europee. – 6. Il diritto alla spiegazione e il diritto ad inferenze ragionevoli. – 7. La proposta di regolamento sull’intelligenza artificiale. – 8. Conclusioni.
Keywords: Artificial intelligence — Machine learning — Inferences — Lack of European laws — GDPR — Right to explanation — Right to reasonable inferences — Artificial Intelligence Act.