Legal Regulation of Civil Liability for Harm Caused by Artificial Intelligence

Abstract

The aim of the study is to systematize international case law on compensation for AI-related harm in order to develop proposals for adapting national legislation to the legal challenges of digitalization. The study employed the following methods: a legal formalism for analysing copyright provisions; a comparative legal method for comparing the approaches of the US and the EU to civil liability for AI-related harm; analysis of case law to identify trends in law enforcement and criteria for assessing damage. The results of the study showed that the case law of the US and the EU demonstrates diverse, but complementary approaches to liability for AI-related harm. In the US, liability is formed mainly through case law, where copyright infringement cases play a key role, as well as cases related to discrimination, opacity of algorithms, and violation of procedural guarantees. It was found that the application of the “fair use” doctrine depends on the transformative nature of the use of content and direct economic harm. In EU practice, the beginning of the establishment of an autonomous approach to liability for AI-related harm is observed. It is proven that the combination of risk-based regulation with civil liability mechanisms creates an effective AI risk management model that can ensure the protection of human rights, while not inhibiting innovative development..