What is AI?

AI or artificial intelligence is technology that allows computers/machines to simulate human intelligence and problem solving. Using algorithms that are modeled after the human brain, AI can “learn” by pumping massive amounts of data through its algorithms. Ideally, the data then allows AI programs to classify or predict outcomes successfully. There are a few different “types” of AI including weak AI (think Alexa or Siri) and strong AI (think Hal 3000 or Ava). According to IBM strong AI is completely theoretical at this point, however, AI researchers are still exploring its development.  

Generative AI (ChatGPT, Dall-E 2, Midjourney, Stable Diffusion) is what most artists and creatives are concerned with. Generative AI uses deep learning models that can take raw data (think every Picasso painting on the internet or every script written by a TV writer) and “learn” to generate statistically probable outcomes, like “new” scripts or an “original” painting. 

The Ethics of AI 

Generative AI like Stable Diffusion can scour the internet in seconds looking for the “raw data”, like images that fit a prompt someone just gave it and generate “new” images. 

 

This image is a composite of everything on the internet that matches the prompt I gave to the Stable Diffusion AI “a painting of a cat on a kitchen table, in the style of Picassos blue period”. The AI then scrapes the internet for images of Picassos paintings and can mix and mash them together to fit the prompt. For many artists this threatens the years, labor, and skill that goes into creating these original works. AI companies argue that images that appear on the internet fall under fair use laws. Fair Use is defined by Stanford University as “In its most general sense, a fair use is any copying of copyrighted material done for a limited and “transformative” purpose, such as to comment upon, criticize, or parody a copyrighted work. Such uses can be done without permission from the copyright owner.” Pamela Samuelson, a professor at Berkeley Law said, “There have been disruptive technologies in the past…” and likened it to the photocopying machine or the home video recorder. The Association of Research Libraries says “as champions of fair use, free speech, and freedom of information, libraries have a stake in maintaining the balance of copyright law so that it is not used to block or restrict access to information. We drafted the principles on AI and copyright in response to efforts to amend copyright law to require licensing schemes for generative AI that could stunt the development of this technology, and undermine its utility to researchers, students, creators, and the public.” The ARL also points to copyright precedents established around machine learning such as Authors Guild v. HathiTrust and upheld in Authors Guild v. Google, where the US Court of Appeals for the Second Circuit held that mass digitization of a large volume of in-copyright books in order to distill and reveal new information about the books was a fair use.  

AI in the Classroom

Academic Dishonesty seems easier than ever to participate in and surely many students will use ChatGPT to at least help write essays that they are unprepared for and unwilling to participate in. The Cornell University Center for Teaching Innovation says that AI detection programs do not provide accurate assessments of student writing. They suggest not using these detection programs at all and instead establishing trusting relationships with the students. The Center for Teaching suggests explicitly communicating expectations in the syllabus and before each assignment, and that where a generative AI is used, it should be credited. Teach for America warns that generative AI can also generate “harmful biases, which reflect racism, sexism, ableism, and other systems of oppression. Last year, Steven T. Piantadosi, a professor at the University of California, Berkeley, uncovered a few of these instances, including responses where ChatGPT said that “only White or Asian males make good scientists” and “that a child’s life shouldn’t be saved if they were an African American male.”  

Using AI

How can we, as artists, use generative AI in an authentic, meaningful, and honest way? ChatGPT can be a great editorial tool for artist statements, project proposals, and other applications if we bring some original writing to the program. I have been experimenting with different ways that ChatGPT can help. I started with asking it to straight up write an artist statement for me, it then asked for the main ideas. After I had given it some concepts that I was interested in, the chatbot went right to working and pumped out a 3-paragraph artist statement. It was mostly fluff, and I could feel the inauthenticity; the writing also lacked the depth and understanding of someone who has worked with clay. ChatGPT was a much better tool when asked to generate an outline, or when asked to simply edit a piece of writing. The method I prefer the most is bringing as much of my own words to the AI as possible. This allows me greater ownership and understanding of the writing that I want. Used as a tool of editing and refining, and not a work around work, ChatGPT and other generative AI programs can be of great use to us. 

Derek Au, the creator of Glazy and a ceramic artist, is working with AI software to create new work based on historical ceramics. There are a ton of interesting ideas within some of the posts on Derek’s website. Here he is using Dall-E 2 and 3 to create what he calls “Alien Vases”, Au says “It wasn’t until the advent of text-to-image generators like DALL·E that my mind was truly blown. Now, one could simply write a text prompt describing the desired image, and results would magically appear. One was no longer limited to specialized datasets, now we were dealing with ALL the images. So far, my primary interest in image generators has been to explore new forms of ceramic objects, primarily vases. I’ve been through many versions of prompts. “A contemporary porcelain vase that has never been seen before.” “A ceramic vase that represents a mathematical concept.” “A modern vase with a drippy glaze.” But none of these prompts produced what I felt were interesting or new forms. As a sci-fi lover, I naturally began asking for vases made by aliens, and my prompts gradually began producing novel forms.” 

Au has a few different projects using AI on his website, such as Virtual Blue and White, which uses 1600 photos to train a StyleGAN3 model.  

Derek Au has also used ChatGPT to “create” new glazes based on 10,000 glazes from Glazy, more specific information can be found here and here.

 

Erin Lynn Smith is a ceramic artist that is using AI as a tool in creating her ceramic sculptures. She says “After feeding hundreds of images of my past work into an AI generator (GAN), Artificial Intelligence created 2 dimensional images inspired by my past work. I then hand built work from these generated images.”