TESTE DE TURING, DIVERGÊNCIAS NA PESQUISA EM I.A. E PENSAMENTO COMPUTACIONAL: UMA LEITURA COMPLEMENTAR DA OBRA O HUMANO MAIS HUMANO, DE BRIAN CHRISTIAN

• DOI: https://doi.org/10.22533/at.ed.8231126280113

• Palavras-chave: Teste de Turing. Inteligência Artificial. Pensamento Computacional. Inteligência Artificial Generativa. Brian Christian

• Keywords: Turing Test; Artificial Intelligence; Computational Thinking; Generative Artificial Intelligence; Brian Christian.

• Abstract: This review article critically examines Brian Christian’s The Most Human Human: What Artificial Intelligence Teaches Us About Being Alive, situating the work within contemporary debates on artificial intelligence, the Turing Test, computational thinking, and generative artificial intelligence. The analysis is grounded in the author’s experience in the Loebner Prize and expands the historical framework to demonstrate that Artificial Intelligence research has always been marked by theoretical divergences, from the Dartmouth Conference to philosophical critiques that question the equivalence between performance, understanding, and consciousness. In this context, the study discusses Turing’s distinction between human and digital computers, Searle’s Chinese Room argument, and recent developments in large language models. It is argued that Christian’s work goes beyond science communication by developing a reflection on language, authenticity, presence, improvisation, error, recognition, and the human condition. As an interpretative update, the article incorporates the analogy of king penguins’ acoustic communication to discuss identity and recognition in complex environments. The study concludes that the book remains relevant because it shifts the question of intelligent machines toward a broader inquiry: which human capacities should be preserved in a context of increasingly sophisticated artificiais systems. As a future research perspective, the article proposes investigating the emulation of the Turing Test through contemporary chatbots and large language models.