THE CIVIL LIABILITY FOR DAMAGES AS A RESULT OF THE DEVELOPMENT, APPLICATION AND USE OF ARTIFICIAL INTELLIGENCE SYSTEMS

AN ANALYSIS OF BILL NO. 2,338/23 (AI REGULATORY FRAMEWORK)

Abstract

This paper critically examines the civil liability framework established by Bill No. 2,338/2023, which aims to create a Legal Framework for Artificial Intelligence in Brazil. Drawing on concrete cases of AI-related harm — including a teenager's suicide attributed to interaction with a chatbot and the unauthorized use of a human voice in advertising generated by AI — the article demonstrates the structural limitations of the bill currently under consideration in the Brazilian Congress. It argues that the legislative choice to refer generically to the Civil Code and the Consumer Protection Code, without creating specific normative parameters, constitutes a regulatory insufficiency. This insufficiency is particularly evident in the absence of adequate treatment for three central problems: algorithmic causation, distributed liability in multi-party systems, and preventive governance based on risk management. The article contrasts the current text with the draft prepared by the Committee of Jurists (CJSUBIA), which proposed a dual civil liability regime grounded in the risk classification of AI systems, and identifies the departure from that structure as a normative setback that weakens victim protection and undermines the preventive function of civil liability.