Why I Support AI-Generated Art
No one asked for my opinion but I couldn’t help but wade into this when I read recently that a bunch of artists on ArtStation changed their profiles in protest of artificial intelligence (AI)-generated art work. It keeps showing up in my news feed and the magazines that I enjoy. There’s the worry about AI-generated work replacing human jobs. There’s a worry about the AI stepping on our copyrights because it was trained on all of our images. There’s just a lot of worry.
I’ll just say it: I’m not that worried.
I work in tech and have for a long time and I’ve seen a lot of trends come and go. For the most part, I have always embraced new technology and the constant advancement of humans in automating some mundane part of our lives. Automation has always been fundamental in making us more efficient or improving my output.
But that said, I am also aware that these automations and efficiencies have made it harder sometimes. People just starting out in the field or those that are unable to learn the harder stuff and move on – well, they also got replaced. So, I just want it to be clear up front that I get it. When machines do a job that humans were doing, some humans are inevitably replaced. And if those humans are not able to move into something new, well, they get screwed.
The fear of being replaced is strong. No one wants to lose their livelihood, their careers, because of changes that result in something else being able to do the job cheaper.
So why in the world would I support AI-generated artwork? Well, it’s just an evolution in technology, like so many things before it, that is likely to cause disruptions and problems but will eventually settle into a tool we use to make our own lives better. I also feel that it’s better to support it and try to help shape it than to wish it would go away.
Wish all you want, it’s not going anywhere. We have been on a constant march towards AI for a long while. It’s used everywhere in so many different things. Of course it’s going to end up in the art world as well. And, someday, it’ll be true AI. The kind of machine that thinks and creates for itself. The kind of machine that creates art for the sake of creating art, like the rest of us. The kind of machine indistinguishable from humanity.
Will there be ethical problems along the way?
Yes. 100%. Algorithms have to be trained on a dataset and most of that data came from publicly available images of copyrighted work. Which means that the algorithm has the potential to reproduce copyrighted work which infringes on the creator. I don’t see this as being much different than us humans though.
If I scour the internet, find a picture I like, and then faithfully reproduce it and post it online as my own work. Well, I did the exact same thing the machine did. People would, hopefully, discover that I copied someone else’s work and didn’t attribute it to them. People would be rightfully mad.
You know, the same way we should be treating the AI-generated artwork. And the people feeding in prompts.
Ethics are still the domain of humanity so it’s really up to us to do the right thing. Today, the machine is simply doing what is asked in the prompt.
Could the algorithm also output a list of artists that influenced the output design? Probably. But a human needs to code that.
Could the human that prompted the AI make sure to tell the world that it was AI-generated? Sure, but this is humanity we’re talking about.
Could we have the AI train on a dataset of images only from artists that agreed to have their art included? I have my doubts that this is feasible. It would require tracking down artists, getting written permission to use their images (which, by the way, we all choose to post online despite the risk of copyright infringement from a human), and then getting enough of a dataset to produce unique results.
Honestly, when you don’t have a lot of data to train your machine learning algorithm on, you won’t get good results. I mean, you’ll get results, but they’ll be skewed. Inaccurate. Wrong. Garbage.
And, in the current conditions, how many artists will agree to allow their images into the dataset?
From where I sit, having seen tech disruptions like this so many times in the last couple decades, I’m not particularly concerned. There will inevitably be some less-than-ethical businesses out there that replace artists with AI-generated art.
But I don’t believe that’s going to be the norm.
I think AI-generated art is going to become a tool in the creator’s toolkit. For artists, these images are going to be a way to ideate faster and result in faster turnarounds for their clients. It’s going to be a way to break out of creative blocks. It’ll be a way to automate and become more efficient. It’s just going to be another tool.
For other types of creators, like say an author, I see this as going to be a way for the author to generate images for their own work. Some authors might accept the images as they are generated – call them “good enough” – and move on with their projects. For others, I see it as a tool to speed up communications with an artist. The author will generate an image, send it to an artist saying, “I want something like this but X, Y, and Z need to be different.”
But what about the unethical business practices and replacing artists with AI?
Even this is grey area. Is it unethical that I create my own flyers and book promotions for my business? I am using Adobe Express to replace hiring a graphic design artist for my business because, frankly, I can’t afford a graphic designer. Is that unethical? Is it wrong of Adobe to create and publish that app?
In the music I make, I leverage an automated mastering service because I can’t afford to hire a real sound engineer. These services tout the usage of AI in their process. Is this unethical of me to try to produce a better quality sound at a price point that I can afford?
We accept AI in many places of our lives in order to make things easier or cheaper. AI-generated artwork is here and it’ll just be a tool in the toolbox. And most of us will probably be better off because of it.
Thanks for stopping by! If you’re interested, I’ve written about ChatGPT – another AI that causes anxiety for people – on my other blog. You can find the post here: The Age of Skynet*.
P.S. All the images used in this post were generated by DeviantArt’s DreamUp service. As a member, I get a certain number of prompts and images per month and I generated a few for this blog post.
P.P.S. In the off chance that you are not aware, the title “Do Androids Dream of Electric Sheep” is the title of a sci-fi novel by Phillip K. Dick that went on to serve as the basis of the movie Bladerunner. So, you know, credit where it’s due.
One response to “Do Androids Dream of Electric Sheep (Or Just Paintings)?”
[…] I don’t often cross-post between my different personalities, but this post crossed into the realm of tech and AI. So, I thought I’d share it here too. You can see the original here: https://landersart.com/2023/01/11/do-androids-dream-of-electric-sheep-or-just-paintings/ […]