Phenomenology of intersubjectivity and cognitive architecture of artificial intelligence

Phenomenological studies of intentional structures of subjective reality are one of the necessary theoretical and methodological resources for building artificial intelligence systems. The article provides a summary of the principles of the classical phenomenology of intersubjectivity and analyzes their application in one of the projects for building a cognitive architecture of general artificial intelligence. In E. Husserl's phenomenology, the intentional structures of transcendental intersubjectivity are given the status of the ontological and semantic primary basis of reality. However, in E. Husserl's concept, these structures are interpreted as immanent interrelations of the field of transcendentally pure consciousness. For Hegel, the structure of the formations of consciousness is formed by the development of the total spiritual culture of mankind. The primary, fundamental fact of the experience of self-awareness is not the cogito, but the experience of being-for-oneself and being-for-another. In my very being, I am dependent on the other, my consciousness at its very center must be mediated by another consciousness. Hegel's approach to the phenomenology of intersubjectivity and its role in the formation of consciousness can be considered as a philosophical justification for the concept of building a cognitive architecture of intersubjectivity in the international project of creating an artificial intelligence system RoboErgoSum. In this project, the objective qualities of objects are revealed through possible actions with these objects. For successful joint actions, interaction agents must be able to coordinate their intentions, plans, goals, and actions. In the same process, there is a perception of oneself as different from the environment, which is the basis of human self-awareness and its robotic model. The project developers insist that in their project robots acquire consciousness and self-awareness, develop the ability to build internal models of the world, to reason and reflect, self-control, goal setting and motivation, achieve an understanding of what is happening, etc. The article concludes that such anthropomorphisms are acceptable only as a metaphorical language of discourse. In fact, we should be talking about modeling higher human mental functions (perceptual, cognitive, affective, predictive) in the cognitive architecture of artificial intelligence. A crucial role in the formation of structures of this kind of integrated intersubjectivity is assigned to the process of artificial intelligence training based on the coordination of robot actions with human actions.