Can the activity of large language models be considered creative work?

Abstract

The article deals with implications of the rapid advancement of Large Language Models (LLM) for the understanding of the creative process using artificial intelligence (AI) and also discusses the new realities in the creative and educational spheres associated with it. The authors propose distinguishing creativity as a formal difference of a product from the already existing one, creativity as a modification of the system of expectations, and finally, creativity as critical and responsible generation of meaning. It seems that LLM are capable of generating formal novelty on their own. In addition, strictly as part of synthetic intelligence and in collaboration with humans, they can participate in changing the horizons of understanding and modifying expectations, as well as in the generation of meaning. However, unlike human, LLM are not capable of autonomous creativity of the second and third levels. Therefore, the effectiveness of LLM is associated with the synergistic effects of human-machine interaction. The authors formulate priority principles of the education system’s transformation in the context of the new digital reality. The current education system is poorly adapted to this new reality, and the skills necessary for “syntellectual activity” of humans and AI in the field of education are mainly formed in students and teachers spontaneously. It is necessary to consciously form and implement a system of principles for using LLM technologies in the educational process. The most important of these principles should be the principle of symbiosis and coevolution of humans and LLM, maintaining their distinctiveness as natural and artificial entities.