Ethical issues in the development of artificial intelligence (AI) have been widely discussed in recent years, not only in the philosophical, but in the scientific liter­ature. This is largely due to the fear of obvious risks and threats that the growing digitalization causes in almost all areas of social life. Ethical research of artificial intelligence (AI) technologies becomes extremely relevant in the context of cre­ating a general AI system, that is, a human-level agent. This task seems difficult since ethics cannot present to engineers a normative system in the form of a cer­tain hierarchical architecture for its computational implementation. In this re­gard, authors of this article see an opportunity to avoid this uncertainty in de­signing an autonomous system of general AI by constructing particular models based on the classifier of activities. This will allow one to algorithmize the nec­essary ethical functions. There are two circumstances that need to be taken into account in further studies of artificial moral agency: (1) the continuality of any algorithmic model of an ethical system and (2) the effectiveness of the “reinforce­ment learning” method in modeling a situation of uncertainty and the mechanism of physiological pain as one of the factors in the formation of moral behavior.

Can the machines that play board games or recognize images only in the comfort of the virtual world be intelligent? To become reliable and convenient assistants to humans, machines need to learn how to act and communicate in the physical reality, just like people do. The authors propose two novel ways of designing and building Artificial General Intelligence (AGI). The first one seeks to unify all participants at any instance of the Turing test -the judge, the machine, the human subject as well as the means of observation instead of building a separating wall. The second one aims to design AGI programs in such a way so that they can move in various environments. The authors of the article thoroughly discuss four areas of interaction for robots with AGI and introduce a new idea of techno-umwelt bridging artificial intelligence with biology in a new way.

In 2016, Hiroaki Kitano proposed that artificial intelligence (AI) will be able to overcome a number of human cognitive limitations that slow down the process of scientific discovery [Kitano 2016 web]. Since then, the odds of AI being awarded the Nobel Prize have been widely discussed, particularly within aca­demic community [Engineering for Research Symposium web 2020]. At the AI Journey 2021 conference, some renowned representatives of four scientific dis­ciplines (physics, mathematics, neurobiology, philosophy) discussed this issue and then co-authored this article [AI 2021 web]. In the first part of our paper, we critically analyze the role of AI technologies in natural science research: how useful they can be for fundamental science, what the potential of AI in natural and exact sciences is, and what principal limitations it has. Another part of our article discusses a counter-question of what science can do for the future re­search into AI. Today, it is impossible to imagine machine learning without lin­ear algebra, physics of materials, and brain research. All this falls under what is now commonly referred to as AI, a general umbrella term [Russel, Norvig 2021]. Thus, having served the birth of AI once, how can physics, mathematics, and neuroscience serve it today?

No abstract