The question, « Can we say that a machine thinks? », has resonated powerfully throughout the ages, fueling a passionate debate that borders on philosophy and technology. René Descartes, one of the most influential thinkers of the 17th century, laid the groundwork for this question by expressing doubts about machines’ capacity to embody true thought. For him, human thought rests on a unique flexibility, an adaptability shaped by experience, intuition, and inner reflection. Machines, however sophisticated, remain confined to their technical limitations and cannot claim the consciousness and introspection that characterize the human mind. Can a machine think? The question of whether a machine can think has always stirred passions and debates for centuries. With the rise of artificial intelligence, this once philosophical question has become a burning issue of current affairs. For pioneers like René Descartes, the answer was unequivocal: no, no machine will ever be able to truly think. Through his reflections, he established concrete reasons justifying his unwavering position. This article explores the foundations of his conviction and how they still resonate today in the context of modern discussions on AI. The Foundations of Thought According to DescartesFor René Descartes, a being worthy of the title of thinker must possess unwavering mental flexibility. This 17th-century philosopher argued that true thought involves the ability to react appropriately to unforeseen situations, an area where machines often fail. Indeed, humans rely on experience, intuition, and introspection to adapt. A machine, however advanced, remains limited by its initial programming. Cartesian reasoning versus the machine In his seminal work, Discourse on the Method, Descartes asserted that human reason functions as a universal instrument, capable of adapting to the diverse circumstances of life. A machine, on the other hand, acts only according to the settings imposed by its creator. Its performance, while sometimes impressive in specific tasks, does not reflect the inner freedom necessary to grasp new and unexpected events. Echoes of the Debate in the 20th Century: This debate, though with ancient roots, experienced a resurgence of interest with the rise of AI in the 20th century. Pioneers like Alan Turing, Marvin Minsky, and John McCarthy worked to establish a discipline aimed at reproducing behaviors that could be described as intelligent. Turing, for example, proposed a famous test, the Turing Test, which posited that a machine could be considered intelligent if it deceived a human during a written exchange. However, even this approach has drawn sharp criticism. The Turing Test: An Illusion of Thought? While appealing, the Turing Test remains merely an illusion. Working with words and sentences does not necessarily imply that the machine can think like a human. Many critics have pointed out that speaking like a human in no way suggests an inner consciousness; the test only assesses mimetic behavior. Descartes would undoubtedly have reminded us that thought depends on introspection and self-awareness, two domains inaccessible to mechanical devices. The Thinking Machine: A Persistent MythRecent developments in AI continue to illustrate this confusion. In 2022, a Google engineer claimed that the LaMDA model displayed consciousness by evoking feelings such as loneliness or sadness. A thorough analysis revealed that these responses were merely echoes of descriptions present in its training data, and that the system generated plausible answers without any real feeling. The Limits of Machine Intelligence Researchers have pointed out that linguistic models, while capable of imitating complex structures, lack lived experience. These systems, such as ChatGPT and Deep Seek, excel at understanding and predicting words, but they possess neither genuine creativity nor moral judgment. Their technical efficiency is an engineering feat, but in no way reflects true consciousness. The emotions perceived in their responses are often merely a human projection onto insensitive algorithms.Despite all these demonstrations, the myth of thinking machines persists, reflecting a blind fascination with technological potential. What do you think? Your thoughts are welcome in the comments!
To read Giorgia Meloni : des images surprenantes en lingerie créées par intelligence artificielle
