Proteção de dados pessoais em sistemas de inteligência artificial generativa: desafios técnicos e jurídicos para o exercício de direitos dos titulares

Abstract

The present study aims to analyze how generative artificial intelligence systems use personal data for the training and functioning of their models and, based on this understanding, investigate the impacts of such processing on data subjects considering the Brazilian General Data Protection Law (LGPD). With this focus, the first chapter presents the history and concept of Generative Artificial Intelligence and examines how this technology relies on personal data for its training and implementation. Next, the second chapter explores how developers and operators of generative artificial intelligence systems can conduct out their activities in compliance with the LGPD, specifically (i) with respect to core principles such as purpose, necessity, and transparency, and (ii) the requirement to identify a legal basis that justifies the processing of personal data. Finally, the last section of the study seeks to answer the following question: once the generative artificial intelligence model has been trained, is it possible for the data subject to request the deletion of unnecessary, excessive, or unlawfully processed personal data under the provisions of the LGPD? The study aims to demonstrate that the right to erasure of personal data from generative artificial intelligence systems faces several challenges stemming from the very nature of these systems, which require new approaches to ensure the rights of data subjects.