Entre Simulação e Compreensão: As Implicações da Crítica de John Searle à Inteligência Artificial Forte

Abstract

The present monograph examines some philosophical implications of John Searle’s critique of the conception of strong artificial intelligence, especially regarding the possibility that machines could possess genuine understanding and consciousness. The objective of the work is to demonstrate that, despite technical advances in the field of artificial intelligence, computational systems do not possess the fundamental elements that characterize the human mind, such as intrinsic intentionality and subjective experience. For this, a theoretical and analytical methodology is adopted, based on a bibliographic review of philosophical and scientific texts, with a focus on the works of John Searle and on contemporary debates about philosophy of mind, cognitive science, and artificial intelligence. The central argument is structured from the Chinese Room experiment, proposed by Searle in 1980, which illustrates the difference between syntactic manipulation of symbols and semantic understanding. The work also analyzes the concept of intentionality and its distinction between intrinsic and derived forms, showing that computers operate only through formal rules and not through understood meanings. Searle’s critique of reductionist materialisms is explored in parallel with the discussion about artificial neural networks, including the challenges faced by these systems due to the absence of an experiential background and of a cultural, social, and personal context— essential characteristics of human cognition. Empirical cases in which artificial intelligence systems fail in interpretative tasks are presented as examples that support the thesis that simulation is not equivalent to understanding. The results indicate that, despite technological advances, artificial intelligences operate through statistical patterns and data correlations, without access to the human Background, a set of non-representational capacities that underlie contextual understanding. It is concluded that consciousness remains an emergent property of biological systems, and that attributing ontological status to machines constitutes a category mistake, with significant ethical and epistemological implications. The study reinforces the need for interdisciplinary approaches to discuss the limits of computational simulation and to value the uniqueness of human cognition.