Gallery Catalog
AI Meets the Art World
Preface
Artificial intelligence lets us create art. Perhaps it also helps us analyze and understand it.
This project is a three-layered experiment: using AI to generate artwork, evaluate that artwork, and then examine those evaluations.
It wasn’t just the art that was built using AI technology. Nearly a dozen different AI applications, many of them released only recently, contributed to the project’s products.
Background
The Large Language Models (LLMs) and ancillary software that power the AI applications are giving us interesting new capabilities. One of these is the ability to convert a text prompt into an image.
There are many examples, often shown in YouTube videos, where people create “art” using one of the many AI text-to-image generators. Mostly, people show that you can create scenes that can’t exist in real life. To me, these computer-derived images are not very interesting. Often, what’s portrayed is dystopian. Such “artworks” aren’t something that I would print, frame and place on my wall.
My reaction was to see if I could create images, generated by the same AI engines, that I would use as wall art. This took me down a rabbit hole.
This document shows how I used many AI programs to create, package, critique, and further analyze images. I call this project the “Gallery Catalog.”
The project consists of three elements: the catalog document (which is a PDF file), a podcast-like critique of the catalog with accompanying images (as a YouTube), and a podcast-like critique of the gallery show after the AI engine has been supplied with additional information (as a YouTube video).
TLDR;
If you’re in a hurry, you can see, and hear, the three primary products by clicking on the links.
Gallery Catalog (a browsable PDF file shown in the ISSUU app as a two-page spread)
Critique 1 (a YouTube video with two art critics discussing the Gallery Catalog)
Critique 2 (a YouTube video of the critics’ comments after they’ve seen the gallery show)
Caveats
This was a progressive learning experience. It builds on experiences using NotebookLM to create audio podcasts.
The idea of creating a gallery catalog was my own invention. I chose the name and location of the gallery to be realistic. No harm was intended with these choices. I hope that Bleecker Street residents are not inconvenienced by people looking for this gallery.
One of the primary goals of this exercise was to show that you can (must?) use a number of AI engines to get deep results. There are many alternative ways to accomplish the results shown here. I happened to use the tools that I understood and which were available to me.
The paid versions of the AI tools were required, in most cases, to achieve these results. In general, these were modest expenses.