Towards the Intelligent Technical Systems with Artificial Consciousness

Analysis of modern trends in the development of artificial neural network technologies, deep machine learning, etc. shows that in the near future, an assault on one of the central and unsolved problems of scientific natural science will inevitably begin all over the world the problem of creating a fundamental theory of consciousness and applying it in practice to create technical systems (TS) with elements of self-awareness, in the image and likeness of humans and animals. If until recently, philosophers and doctors paid attention to the problems of consciousness, then in the last two decades, in connection with the mass introduction of "intelligent" information and robotic technologies into practice, including medical technologies, serious engineering publications began to appear on the issues of the existence of consciousness in living nature and the tasks of its reproduction in TS. This review collects and analyzes the main data known in biology, medicine and natural science in general on the phenomenon of consciousness. Arguments are given in favor of the possibility of an engineering and technical formulation of the problem of creating intelligent TS that are separately aware of themselves and the world around them. A number of considerations are also put forward as to what basic principles may underlie such engineering and technical hardware and software solutions. A simple working engineering and technical definition of consciousness is proposed as a high-level control mode of a multisensory mobile system designed to act autonomously in a changing external environment, guided by personal goals and meanings. Based on well-known examples from wildlife, the regularity of the evolutionary emergence of consciousness in animals is substantiated both in terms of phylogenesis and early ontogenesis. The indispensable role of the endocrine system in control in the animal world is shown, as well as a concept for its implementation by technical means in conjunction with neural networks is proposed.