Abstract

The capacity of generative AI (GAI) has improved markedly. Today, virtually anyone can use publicly available tools with powerful generative capabilities – such as ChatGPT – which has led to an explosion of AI-generated content, some of which purports to be “art.” As a result, debates have emerged over whether AI-generated works belong in the art world. This essay examines GAI-generated works through the lens of four kinds of theories on how to define art: 1) artistic expression, 2) aesthetic experience, 3) historical or institutional context, and lastly, 4) artistic intention. The conclusion is that while GAI output can theoretically meet the criteria of the first three kinds of theories, the decisive factor appears to be the user’s degree of control over the creative process, and by extension, the end result. In other words, a sufficient degree of intention is required for GAI output to constitute art, which remains difficult, if not impossible, to achieve with currently available tools.