COGNITIVE DEPENDENCE ON AI ASSISTANTS IN ARTS EDUCATION: STRATEGIES FOR PRESERVING CREATIVE SUBJECTIVITY

Abstract

This article examines the phenomenon of cognitive dependence on generative Al tools in the context of training students in arts disciplines. The study is motivated by the rapid increase in accessibility of image-, music-, and text-generation systems (e.g., Midjourney, DALL E, Suno, ChatGPT) and by the attendant risk that authorial subjectivity will be displaced from the creative process: learners increasingly delegate not only auxiliary operations but also the generation of artistic ideas, selection of figurative language, and compositional decisions to technological agents. Drawing on cognitive load theory (Sweller et al.), the concept of cognitive offloading (Risko & Gilbert), and Bloom’s taxonomy of educational objectives, the article systematises the mechanisms by which generative Al intercepts the cognitive operations that directly shape artistic thinking—primarily the lower cognitive levels that serve as the foundation for analysis, evaluation, and creative production. The paper reviews experiences of integrating Al tools into professional arts training and argues for the necessity of a systematic pedagogical response to the challenge of dependence. Two groups of pedagogical strategies for preserving creative subjectivity are proposed: authorial-identification strategies (Socratic questioning, verbalisation of the creative process, artist’s journal) and strategies that treat Al as an object of critical interrogation (verification of Al-generated outputs by professional criteria; comparative analysis of “own concept versus Al variant”). A three level conceptual model of interaction—learner, Al assistant, and educator—is advanced, with clearly defined points of design, diagnostic, and developmental pedagogical intervention in the arts learning process. It is argued that preserving creative subjectivity in the era of widely available generative Al requires not technological prohibition but systematic pedagogical management of learner-technology interaction at every level of artistic learning activity. The findings can inform curriculum design and methodological resources for arts institutions and provide a theoretical basis for future empirical research in this field.