O tempo do encontro entre o direito e a inteligência artificial: reflexões fenomenológicas a partir do pensamento sem corrimão de Hannah Arendt

Abstract

This doctoral dissertation, developed within the Group for Studies, Research, and Extension in Public Policies, the Right to Health, and Bioethics (Biogepe), affiliated with the Graduate Program (Master’s and Doctorate) of the Faculdade de Direito de Vitória (PPGD/FDV), investigates the impacts of deep learning–based artificial intelligence on the human condition of work and the limitations of Brazilian law in providing adequate social protection in the context of advanced automation. The connection to the graduate program and research group is justified by the centrality of the topic to contemporary debates on public policies, fundamental rights, bioethics, and new forms of structural vulnerability produced by technology. The title of the dissertation indicates the adoption of Hannah Arendt’s political phenomenology as an analytical grammar for understanding artificial intelligence not merely as a technological instrument, but as a phenomenon capable of affecting fundamental dimensions of the human condition—labor, work, and action—with significant legal implications. The central problem examined is whether modern legal categories, grounded in instrumental rationality and a traditional conception of work, are sufficient to address the social, normative, and anthropological effects of automation. The hypothesis advanced is that artificial intelligence operates as a structuring force of social life, producing a gap in legal protection that Brazilian law, in its current configuration, is unable to adequately address. The methodology is qualitative and critical-phenomenological in nature, inspired by Arendt’s notion of “thinking without banisters.” It combines specialized bibliographical analysis in political philosophy, legal theory, and artificial intelligence studies with documentary research, including the examination of recent empirical data on automation and employment, as well as the analysis of bills, regulatory documents, and normative initiatives related to artificial intelligence and the protection of work. The first chapter situates the historical centrality of work in modernity and the emergence of automation. The second chapter presents the Arendtian grammar of the human condition, distinguishing labor, work, and action. The third chapter analyzes deep learning– based artificial intelligence as a technical and social phenomenon, examining its empirical impacts on employment and labor dignity. The fourth chapter investigates the Brazilian legal response to automation, with particular attention to the constitutional provision of Article 7, item XXVII, and existing legislative proposals, highlighting the gap in social protection. The fifth chapter proposes a conceptual category to describe the emergence of a subject whose cognition and action are progressively mediated or replaced by technology, discussing its legal implications. The dissertation concludes that artificial intelligence simultaneously places labor, work, and action at risk, requiring a legal response that goes beyond merely instrumental solutions. As a contribution, the study offers a critical phenomenological reading capable of informing future public policies and normative debates on automation and social protection