¿Son las computadoras agentes inteligentes capaces de conocimiento?

The aim of this paper is to study the philosophical foundations of the programming architecture in two Artificial Intelligence sys-tems (AlphaGo and Hide and Seek). The problem elucidates the epistemological distinction of the concepts “knowledge,” “intuition” and “abduc-tion,” in order to define whether the successful fulfillment of a programmed metric, by a com-puter, is a sufficient condition to attribute intelli-gent behavior to it. Through the analysis of both examples, two issues are shown: i) the successful fulfillment of a programmed objective offers new knowledge resources, ii) such knowledge depends on the execution of a program whose processing is developed by an AI and, therefore, the opera-tions exceed human intellectual capabilities. The conclusions point to the fact that computers are special hypothesis-testing knowledge resources.