From infant robots to recursive models: historical consciousness and the exhausting of exhaustion by generative AIs

Following a thread of thought that, since the beginning of the nineteenth century, is running through the history of mathematical theory and information science (at least since A. Turing’s imitation game), it is possible to individuate a conceptual constant which has been predominating in most scientific and philosophical discourses about the relation between human and machinic intelligence. This constant is a particular (self) conception of human consciousness, and of its relation with autonomous technologies, based on the presupposition of the machine imitating, competing, finally overcoming (or not) the human. In order to explore this constant and to put it under a critical light, the theoretical point of departure of this essay will be Sylvia Wynter’s proposition that human mental states, and ideas of what it means to be human, are cultural constructions, and are therefore to be investigated through a combination of scientific and cultural research. Another important reference will be the work of Neda Atanasoski and Kalindi Vora, and their discussion of technology as fundamental for the definition of what it means to be an autonomous human subject. Following the input of these two theoretical references, the paper will reveal the presence of the culturally specific idea of imitation (ie an imitative paradigm of human–machine relationality developed by modern Western thought) in a particular theoretical toy-model of AGI. This toy model is Kanzi, a speculative prototype through which the philosopher Reza Negarestani tries in fact to reverse the paradigm and to reconsider the human itself as a toy-like AGI. Many applications of generative automation (or Generative AIs) in fact exemplify today the necessity for a continuous re-generation of the same paradigm, contributing to