Filosofia da inteligência artificial: Antropocentrismo vs. Pensamento maquínico

Artificial intelligence (AI) systems now match or surpass human performance across an expanding set of cognitive tasks, prompting a reassessment of long-standing assumptions about consciousness and moral status. This article interrogates the opposition between anthropocentrism, which treats rational agency and first-person phenomenality as uniquely human, and machine thinking, which frames cognition as a substrate-independent computational process. Using conceptual, comparative, and critical methods, and following PRISMA‑style screening and thematic coding, we analysed 89 journal articles indexed in Scopus and Web of Science, alongside canonical texts by Descartes, Kant, Turing, Searle, and Dennett. Three patterns emerge. First, contemporary large‑language models reinforce anthropocentric intuitions by mimicking phenomenality. Second, formal arguments for machine consciousness increasingly draw on predictive‑processing, Global Workspace, and Integrated Information accounts to challenge species boundaries via substrate-neutral criteria. Third, hybrid ethical frameworks combining anthropocentric precaution with machine-oriented functional indicators offer the most coherent route for science and policy. Focusing on Ukraine, a rapidly growing AI hub, we demonstrated how this hybrid approach can inform national strategies that align with the EU AI Act while respecting local narratives of human dignity and security needs. The study clarifies conceptual gaps in current debates and outlines a philosophically grounded roadmap for inclusive, risk-sensitive AI governance. From the standpoint of the methodology of meta-anthropology of AI, it is shown that in the future a human will be able to communicate with AI not simply as a device that he owns, but as a subject, an other, who has his own existence and the right to freedom.