Will a human always outsmart a computer? An essay

Abstract

The title question of the paper has its empirical origin in the form of an individual’s existential experience arising from the personal use of a computer, which we attempt to describe in the first section. The rest of the entire paper can be understood as a philosophical essay answering the question posed. First the connection between the main problem of the article and its “premonition” by mankind, which was expressed in the form of ancient myths and legends, is briefly suggested. After shortly discussing the problems that early considerations of AI focused on, i.e. whether machines can think at all, we move on to reformulate our title question, about the possibility of outsmarting AI. This outsmarting will be understood by us in a rather limited way as to prevent a machine from completing its implemented task. To achieve this objective, after softly clarifying the basic terms, an analogy is built between the “outsmarting” of a machine by a human (the target domain) and the playing of a mathematical game between two players (the base domain), where this outsmarting is assigned a “winning strategy” in the certain game. This mathematical model is formed by games similar to Banach-Mazur games. The strict theorems of such games are then proved and applied to the target of the analogy. We then draw conclusions and look for counter-examples to our findings. The answer to the title question posed is negative, and it is not clear how far it should be taken seriously.