Frugal AI-Generated Art: A Poetic Form and Ethical Choices

Abstract

Generative artificial intelligence has transformed the production of visual art and design, while also raising concerns about energy consumption, copyright, visual homogenization, and the weakening of creators’ subjectivity. In response to the tendency to measure artistic value by model scale, computing power, output precision, and production volume, this paper proposes the concept of “frugal AI-generated art.” In this approach, creators deliberately limit model scale, data volume, generation frequency, output precision, and visual complexity. Rather than treating these limits as technical shortcomings, they turn resource constraints into artistic forms and expressions of value. Frugal AI-generated art is therefore not a passive downgrade or a rejection of technological development. Instead, it uses deliberate limitations to stimulate creativity and to moderately reconstruct the relationship between human creators and computational systems. By foregrounding the creative potential of constraint, this concept offers an alternative to the assumption that greater technical capacity and higher output necessarily produce greater artistic value. It invites consideration of how choices about resources and visual form can shape both the creative process and the values expressed through AI-generated art.