Intellectual Property in Language Models: Challenges of Ownership in the Integration of Multiple Databases

Abstract

This work is the result of the project Innovative Solution for Generative Artificial Intelligence for Automated Responses, developed within the research group Centro de Competências em Soluções Livres at IFRN (CCSL-IFRN), and addresses the legal challenges related to intellectual property ownership and data protection in the development of language models based on Artificial Intelligence (AI). The study of this topic is relevant due to the insufficiency of current legislation in dealing with issues such as authorship, use of sensitive data, and liability. The objective of this work is to analyze these challenges and propose guidelines for a regulatory framework that balances innovation and the protection of rights. The methodology consists of qualitative research, using a hypothetical-deductive approach, employing bibliographic, documentary, and jurisprudential analysis. Based on the analysis conducted, it is concluded that a specific regulatory framework is necessary to address the peculiarities of language models, ensuring clarity in defining ownership and respecting the rights involved.