On the Problematic of Modeling Consciousness in the Tasks of Artificial General Intelligence
Artificial societies January 1, 2026 DOI: 10.18254/s207751800038254-8 (opens in new tab)
Summary
AI-generated from the abstractAchieving general (strong) artificial intelligence may require implementing subsystems for consciousness, subject and subjectivity, self-awareness, intentionality, and homeostasis as constructive cybernetic objects. Connectionist models like large language models, relying solely on self-organization and emergence, may be insufficient. The history of Turing tests indicates a need for a set of metrics to objectively judge when a cybernetic system reaches the level of general artificial intelligence.
Study at a glance
| Characteristics | Theoretical or philosophical paper Peer reviewed |
|---|---|
| Key finding | Argues that without implementing consciousness, subject, self-awareness, intentionality, and homeostasis, it is impossible to achieve general (strong) artificial intelligence, and that a set of metrics is needed to evaluate such systems. |
Abstract
Discussions about general (strong) artificial intelligence, the fundamental possibility of achieving it, as well as whether it is possible to implement it within the framework of connectionist models alone, such as Large Language Models solely due to self-organization processes (i.e. based on the ideas of emergence), remain one of the hottest today. On the other hand, today the question remains about the list of subsystems and mechanisms (process models), without which it would be impossible to characterize the designed system as a system of general (strong) artificial intelligence. Understanding what (which subsystems) should be implemented in the case of solving the problem of creating a system of general (strong) artificial intelligence is crucial here. In our opinion, without the realization of such phenomena as consciousness, subject and subjectivity, self-awareness, intentionality, as well as such properties of complex systems as homeostasis as constructive objects of cybernetics, it is impossible to talk about consciousness and conscious activity of certain artificial systems. In addition, the history of the development of the entire variety of tests (for example, combined into the group "Turing tests") shows the need to create a set of metrics that would allow making objective judgments about achieving the level of general (strong) artificial intelligence in certain instances of cybernetic systems.