As obras artísticas geradas pela inteligência artificial: considerações e controvérsias

• Ano de defesa: 2019
• Autor(a) principal: Ferro, Vanessa da Silva
• Orientador(a): Dias, José Carlos Vaz e
• Banca de defesa: Anna, Leonardo da Silva Sant , Branco Júnior, Sérgio Vieira
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade do Estado do Rio de Janeiro
• Programa de Pós-Graduação: Programa de Pós-Graduação em Direito
• Departamento: Centro de Ciências Sociais::Faculdade de Direito
• País: BR
• Palavras-chave em Português: Inteligência artificial; Direitos Autorais; Requisitos de Proteção; Aprendizado de Máquinas; Autoria; Contributo Mínimo; Domínio Público; Fundamentos dos Direitos Autorais
• Palavras-chave em Inglês: Artificial intelligence; Copyright; Protection Requirements; Machine Learning; Authorship; Originality; Public domain; The Foundations of Copyright
• Área do conhecimento CNPq: CNPQ::CIENCIAS SOCIAIS APLICADAS::DIREITO
• Link de acesso: http://www.bdtd.uerj.br/handle/1/9485

It is believed that artificial intelligence (AI) has autonomy as well as the ability to learn, emulating human cognition and, in this way, can be able to produce artistic works without explicit prior programming. With this assertion, one can immediately identify conflicts with the copyright law. In Brazil, copyright gives authors legal protection in relation to their artistic productions. However, human authorship is a prerequisite for proprietary protection in the country. Therefore, works generated by artificial intelligence currently do not match with protected intellectual works of authorship. On the other hand, there is evidence that people, in their daily lives, perceive the utility and heritage or axiological value (scientific, cultural and social) of the works generated by the AI. Such works are perceived as apt to provide utility and/or meet the needs of people, in such a way that they can be characterized as appropriable legal assets. Considering that the legal system prevailing in Brazil is based on the appropriation of assets, including intellectual property, the debate on the works generated by the AI is relevant to address the problem related to the potential legal protection of such works. At the same time, an unfolding of this discussion leads to research related to authorship, addressing certain legal fictions in Brazil, the United States, the European Union and the United Kingdom. Within this scenario, this essay is relevant because it analyzes the potential attribution of proprietary rights to third parties other than authors, aiming at stimulating the disclosure of the works generated by the AI. In this scenario, the present study will address the potential paradigm shift in the artistic works generated by AI, as well as the methods and some machine learning techniques, which grant AI the autonomy to produce artistic works. A brief overview of the copyright system will be drawn, addressing its modern origin, with proprietary bases, centered in the granting of privileges focused on the reproduction of copies until the paradigm shift that historically marked the bias owner of the copyright oriented to the figure of the author. The requirements for the protection of intellectual works, such as human authorship and originality, in addition to the following main theories justifying copyright (based on the methodology adopted by Professor William Fisher III), will be discussed: (a) on the one hand, the justice (or labor) and personality theories, under a conception of natural rights; and (b) on the other hand, the welfare and cultural theories, under a utilitarian perspective. In this sense, we will investigate the existence of justificatory grounds for proprietary protection of the works generated by the AI in light of the so-called Fisher Methodology