It's important to keep in mind that these art communities represent one very narrow kind of art, one that is more focused on illustration and graphical style, and most often not the kind of art that major art institutions work with. In the past 5 years, AI-generated art has been accepted and incorporated in the contemporary art world, with major works produced by luminaries such as Pierre Huyghe (of Ideal), Trevor Paglen (Bloom), and Hito Steyerl (Power Plants). For these artists, the graphical potency of the generated images is usually only one factor of the work, which also centers on how the artists use the technology, its origins, history, and broader uses, and what it might imply.

For this kind of artwork, the kind made by artists who feel threatened by AI, I’ve been thinking for a while that a computer could likely do it better. Intricate detail, repetition, sampling and recombination of popular styles, optimization for aesthetic popularity and mass appeal — these practices are all squarely in the domain of automation, data science, and marketing. We might call it the Netflix-ification of illustration art, but it’s quite different, because the tools for making are available to many. As an arts educator and artist, I welcome these new tools, as they allow easy experimentation for newcomers, but also as they reveal that art based primarily in deployment of technical skill and eye-popping graphics is missing something vital.