The Subsystem of Emotional Intelligence as a Necessary Condition for the Implementation of Learning and Self-Learning Processes in the Tasks of General Artificial Intelligence

The implementation of strong (general) artificial intelligence is one of the most discussed topics today. However, today the main discourse is, as a rule, in the limited field of connectionist methods alone. In our opinion, the creation of a system capable of effective operation in constantly changing environmental conditions is possible only within the framework of creating hybrid models that combine not only connectionist, symbolic methods, but also humans. A necessary condition for building such systems is also the implementation of the subsystem of emotional intelligence. Unfortunately, the topic of emotionality when discussing AI issues today is often reduced to the need to recognize human emotions or as part of a more general task related to emulating a certain emotional response during the interaction of some instance of an AI system and a person. In our opinion, this understanding of the issue is associated with a simplified understanding of the role of emotions and a narrowing of their importance to auxiliary tools necessary for a better understanding of the participants in the dialogue. In our opinion, without the implementation of a full-fledged subsystem of emotional intelligence, neither the implementation of learning and self-learning processes, nor effective decision-making in an uncertain environment is possible. The purpose of this work is to search for and rethink the foundations that could serve for a philosophical understanding of this issue, as well as those foundations that could be useful in designing AI systems in practical terms.