A evolução do projeto computacional para uma inteligência artificial e as novas perspectivas oferecidas pelos avanços da cognição enativa

Abstract

The development in the Artificial Intelligence (AI) field brought great improvement to the resolution of computationally hard problems. However, the early goal of AI, of implementing a genuine intelligence in an artificial way, was not achieved. Therefore most of the field’s researchers changed their research’s focus to problem resolution, quitting the goal of an artificial intelligence. To keep pursuing that goal others researchers started questioning the philosophical principles of AI and they took new routes. Some of them realized that the process of intelligence is not purely mental. According to this new view the body has a fundamental role in the cognitive process. From this point of view, aiming to overcome traditional obstacles of AI, the Embodied AI emerged. It has a strong tendency to implementation in robots, to develop a better design of the body. Despite having achieved improvements over the problem of traditional AI, Embodied AI started to present its own limitations. Then came the idea that a genuinely intelligent agent must formulate its own problems from the perception of reality, constructed in terms of their sensory-motor apparatus. In other words, the genuine intelligence is linked to the agent’s autonomy. The Enactive AI appeared influenced by biological studies about autonomy. In this AI field the goal is to build an autonomous artificial agent. This paper reports the trajectory of AI since its foundation, through strand of Embodied AI, and pointing to a possible new paradigm of Enactive AI. In addition, we analyse and discuss the processes that led researchers to question the philosophical basis of IA and formulate new concepts about what intelligence is.