Update as of December 5, 2023 – Interesting new tool being utilized by artists to “poison” their work by adding elements to defeat being used for AI – https://www.technologyreview.com/2023/10/23/1082189/data-poisoning-artists-fight-generative-ai/
The next few posts will revolve around artificial intelligence (AI), beginning with the use of AI and art. AI in general is a hot topic because it’s a potential game changer in education, creation, research, and even blog generation. And yes, there’s a human behind this post.
I’m going to start with AI and art creation because, as a former art teacher, I’m still following a lot of sites and news stories that have to do with art, and it’s a topic that is causing curiosity and concern and outright anger.
Incidentally, I’ve seen mention of the terms synthography – traditional referring to diagnostic imaging in nuclear tests – now suggested as a term for image production created with AI. Synthetic Media is another phrase being mentioned, referring to video, audio, images, or text where AI is used for all or part of the creative process.
When is Art Original?
Many people have tinkered – or outright created – art on digital platforms. From basic Microsoft Paint to Adobe Creative Suite, lots of options exist. And they are fun. When you use these, you usually start with paintbrush and drawing tools, but then explore pre-made filters and brushes and special effects. You are still making choices as to composition, colour, effects, etc. so it’s your work. But what happens when you take a photograph and then use those tools to edit out imperfections or add little extras? Is there a line? Does it depend on who took or who owns the photograph? What about artists like Roy Lichtenstein or Andy Warhol, who manipulated the work of others by adding colour or enlarging and adding patterns?
If you buy into the philosophy of Kirby Ferguson’s Everything is a Remix, then nothing is original. Nothing? I don’t believe that’s true, but there’s something to be said for being inspired by what comes before. Sir Isaac Newton once said, “If I have seen further, it is by standing on the shoulders of giants.” I’m not quite there – I strongly believe that there is originality and creativity and innovation happening all the time. Technologies are being invented all the time. So are methodologies, theories, etc.
Plagiarism in Art
This brings up the need for a discussion about copyright and plagiarism. That’s a whole topic on its own and I’ll link up some thoughts later on. What sticks in out in my mind is the difference between appropriation and inspiration. Or transformation. If a work is substantially transformed, then can someone sue for copyright infringement? It’s not always clear. My Sweet Lord (The Beatles’ George Harrison) was clearly copied from He’s So Fine (The Chiffons) and Harrison was fined $587K for what he described as subconscious plagiarism. There are also famous cases in the fine art world, such as Jeff Koon’s copy of work by Franck Davidovci. And a recent case of Jingna Zhang vs Jeff Dieschburg. I could go on. And on.
While there are lots of examples, often it’s not clear-cut. How much transformation or alternation is needed to label a work as original? What is considered enough of a transformation is subjective. It might also depend on what defines the essence of the piece. Fair Dealing or Fair Use can also effect how the use of the work is viewed. (Canadian copyright differs from the U.S.)
So this is the background behind the current climate and has a large part in how artists are viewing the use of AI to generate work – from their own art.
Lawsuit
Artists are acting. There is a class-action lawsuit for artists whose work has been used to train on Midjourney & Stable Diffusion. There are two Google Docs you can check, for Midjourney Artist Styles and Stable Diffusion prompts. Go to this Twitter thread to learn more: https://twitter.com/Kelly_McKernan/status/1615066246591938560 and also includes interesting chatter about this.
My Thoughts
I’m leaning towards there being a difference between training on art and AI generators like ChatGPT. I’ve played with ChatGPT and a few others and so far it seems that they are not outright copying literary writing, just kind of mixing and matching what’s out on the web. And also making up facts and citations. I still need to play more with the different kinds of prompts. But for art? If I painted something and found an obvious part of my painting as part of someone’s uncredited AI-generated art, and even worse, for sale as digital art, I would not be happy. Same with anything arts-based. I’m going to follow the class-action. (Note: it’s not up on the law site as of Jan. 17 2023)
ChatGPT Summary of Issue
Because why not?
- Artists may feel that their work is being used without their consent.
- They may also be concerned about the loss of control over how their work is used and perceived.
- There may be financial concerns, as the use of their work in AI training may generate profits for the company or organization using it without compensating the artist.
- They may also be concerned about the potential loss of meaning or context in their work when it is used in AI training.
- Some artists may be worried that AI-generated art will devalue their own work.
- Some may be worried about the potential for their work to be misinterpreted or misrepresented by the AI.
And then to be fair, I put in: Are there any positive aspects to art being put into AI training?
And the cursor sat there for quite a while, thinking. Then it responded with:
There can be several positive aspects to art being put into AI training. One of the most notable is that it can help to create new and unique forms of art that may not have been possible before. For example, AI-generated art can be used to create new styles of painting, sculpture, or music that blend elements of different art forms. Additionally, AI-generated art can be used to create new forms of storytelling and other forms of creative expression.
Another positive aspect is that AI-generated art can be used to help preserve and archive art. For example, by training AI models on existing artworks, it can help to preserve images, sculptures and other forms of art that might otherwise be lost over time.
AI-generated art can also be used to help educate people about art, by providing new ways to explore and understand different styles and forms of art. Additionally, it can be used to help create new forms of interactive art and create new ways for people to engage with art.
Lastly, it can be a source of inspiration for the artists themselves, by giving them new ways to think about their craft and new perspectives to create from.
So you decide. Is copying (large) parts of someone else’s art infringing on copyright laws? Plagiarism? Where do we draw the line?