Image by Midjourney
AI is such a contentious topic right now that I feel as though it would be good to be transparent about how I use AI with the works connected to Naja Tau, and outline why I use it the way I do. Because I, personally, am not interested in reading AI creative writing. So I wouldn't blame you if you were also not interested. So if an author is copying and pasting medium to large chunks of text from AI, I don't want to read it. I wouldn't mind reading non-fiction articles that clearly state that the article was copied and pasted from AI (it would be a big bonus if a human expert had read the article and validated what it was saying). But I'm not interested in reading stories and fantasies from a non-biological entity.
I write because I have something to say. I don't want a piece of software to override what I want to say and how I want to say it. However, I will use AI to tell me about facts, rules (legal, grammatical, formatting, and structural), and I will sometimes copy and paste the descriptions of my books into AI and ask the software to analyze it and give recommendations for improvement. Sometimes, the AI will tell me I need to add a stronger hook, or a call to action. And I'll rewrite the description and ask it to analyze the changes I made. I'll keep doing that until I feel good about the description. I did this for the most recent description of Naja Tau's Dream Diary 2.
Sometimes, the AI will suggest that words in a sentence swap place or it will recommend the use of a more active voice rather than a passive voice. And in those cases, I might accept the exact wording of a recommended change. I did this with the description of Satyr Plays 2.
I have also asked AI to generate lists of things. Lists of words that rhyme. Lists of words that mean x-and-such. Lists of fun, tropical recipes (for Satyr Plays 1).
I have also copied and pasted legal disclaimers about copyright, and changing names, locations, and places, and the fragility of human memory.
I'm not interested in copying and pasting actual passages of fiction or creative non-fiction from AI. I don't ever plan to do that. Could some weird advancements in AI change the way I do things? I guess so. You never know. But this is what I've been doing with AI since it started to get good in about 2022.
Of course, there's also the matter of using AI-generated art. Under most of my square images in my blog, I say, "Image by Midjourney." Midjourney is an AI art generator. I also use my Midjourney subscription for a lot of YouTube thumbnails. I have also used Midjourney to generate the cover images for Satyr Plays 1 & 2, and Megachurch Versus Tattoo Studio.
Although I use AI art, I use AI art because I'm poor. Would I rather pay a human artist that can crank out nice art really, really fast? I would. It just leaves one with a warmer feeling to know that human hands and a human mind created art. But before Midjourney, I wanted to have art on the blog, and it's just too expensive and time consuming to do. There just weren't any images at all on my blog--or I might have occasionally grabbed them from Morguefile.com or Unsplash.com.
I like having AI images though, because it creates a nice separator between blog entries, and it creates a visual anchor for remembering whether or not you've already read a post.
I think that the arguments for whether or not using AI-generated art is acceptable change depending on whether or not the AI is grabbing artists' images and slightly changing them and reselling them as its own, or if it's thoroughly remixing a bunch of artists' images and creating something new. Because this is how we evaluate whether or not people are doing something unethical with other people's art. If a person grabs a JPG of my drawing of a macaw and puts it on merchandise and sells it (which has happened to me), that's stealing. But if a person looks at my drawing of a macaw on the internet, gets inspired by that image, and looks at other images of macaws, and blends them all together into something new, that's just how the human brain works. That's how new art gets made. Humans see images that inspire them, then blend it all up in their brains and create something different out of everything they've seen. Why are we holding our robots to a higher standard than we could hold a person to in this case?
I was concerned that Midjourney was just copying artists' images, so I started doing reverse image searches on my YouTube thumbnails. So far, there aren't any similar images to the ones Midjourney generates for me.
Also, if Midjourney uses hundreds of millions of images to train on, and we decided that all of the people who made those images should get paid for that, do we really think that every time Midjourney generates an image, a person should get a hundred millionth of ten cents (assuming that an image Midjourney makes will generate ten cents for the company)? I guess that would add up. And maybe one day, those artists will be identified and can sue for that. Sure. But there's no way to hold humans accountable for doing this with far fewer images in their "database" (brain). Should humans be held accountable in that way if they can be? I don't think so.
So that's where I'm at right now with using AI for the stuff connected to Naja Tau.

No comments:
Post a Comment