Subject-like Qualities of AI: “Docking” Humans and Non-humans

Abstract

The article discusses the problem of subject-like qualities of functioning of arti­ficial neural networks (ANNs). The concept of non-humans refers to self-orga­nizing intelligent devices on non-biological carriers. The main goal of the arti­cle is to compare human subjectivity and qualities demonstrated by ANNs.
The author records the transformation of subjectivity, caused by both the hyper-interconnectedness of the individual with the digital world and the functioning of the ANN. In view, firstly, are the features of traditionally understood subjec­tivity with an emphasis on reflexivity. Secondly, there are the phenomenon of the informational “global workspace”, significant for the modern understand­ing of consciousness, and those features of ANNs, that lead to the hypothetical recognition of their subjectivity. Thirdly, the author conducts the analysis of hal­lucinations of ANNs. Fourthly, the author explores the meaning of the prompt in the context of the “docking” and divergence of humans and non-humans. The research methodology involves the comparative analysis, socio-humanitar­ian interpretation of the functioning of the ANNs and a subject-oriented ap­proach in the context of ergonomic interaction with AI. It is concluded that the signs of subjectivity of ANNs include their autonomous functioning, the ability to self-learn, the possibility and capability, reminiscent of reflection, not only to collect information, but also to transmit the hidden layer of the obtained values to itself at the next step. The current status of human subjectivity acquires the qua­lities of an actant, implementing a sequence of functions necessary to carry out life on the Internet, thereby “docking” humans and non-humans.