Nearly three years ago, I made a post about my initial thoughts on the proliferation of generative AI technology in the context of creating art. Since then, the technology hasn’t gone anywhere—it’s evolved the same way all technology does, but at this point, it seems pretty clear to be a permanent fixture in the creative landscape, at least for the immediate future.
What got me thinking about this again was this piece of fanart that I recently made.
This image is a composite consisting of six separate stock illustrations: one for the background, one for each character, and one for the eyes of the second character from left.
Here is the canvas playback, where you can see how I composited the images together and the changes I made. Warning: there are some parts with strongly flashing black and white areas due to cropping of the canvas. If you have epilepsy or similar photosensitive conditions, exercise caution.
For those who can’t or don’t want to watch the playback, here is a summary of the changes made to the stock images I used:
- Background: cropping/aspect ratio; otherwise unchanged
- Far left character: original pose was standing, I changed the pose to sitting; made adjustment to hand/arm position, minor changes to clothing details, changed facial expression and direction of eyes
- Second from left character: changed line thickness to match the rest of the image, redid the feet, minor adjustments to pose, changed the facial expression, replaced eyes with another stock image
- Third from left character: original pose was standing neutrally, I changed it to leaning from behind the far right character; minor changes to clothing details
- Far right character: minor changes to clothing details, changed facial expression
Anyway, while I was doing all this, it occurred to me that my overall process in creating this piece is fundamentally similar to how a generative AI tool creates an image. The end result is my own creative vision, but the component pieces are all stock assets licensed from Shutterstock. Legally, I have the right to use them for personal use, but that still doesn’t make them my original work. So, what makes this image any different from an AI-generated image?
As far as I can tell, any greater value ascribed to my image comes solely from the fact that I, a human, made it without assistance from an AI, and the stock assets I used were also made by humans, and not AI.
There are people who get hoppin’ mad about art generated by AI and the nature of how AI art tools are trained, namely by using existing artwork as a dataset. Personally, I just don’t get it. I mean, I understand the logic of why people might get so upset about those things, but those reasons aren’t things I can relate to. I don’t see AI generated art as some great evil and/or insult, because by and large, it doesn’t have the capacity to replace human creative direction. It doesn’t bother me if an AI is trained on my work, because machine learning is still learning. If I look at a hundred drawings of a rabbit and then draw a rabbit based on them, how is that different from an AI tool doing the same thing? Just like in the case of the fanart featured in this post, the only thing I can see is the human element.
To further explore people’s opinions about AI art tools and how they intersect with human artists, I recommend this Reddit thread from a year ago.
For more on why we tend to prioritize and value the human element as a general principle, check out this Reddit post exploring the question of why we consider human life valuable by default.
Philosophy, as interesting as it is, is not a core focus of this blog, so I will leave you with those two links rather than pontificating further.
Looking back at my original post about AI art, my opinions overall haven’t changed significantly. If you missed the original post, those were as follows:
- I don’t have a problem with AI art tools when they are used in support of a specific creative goal the user has.
- I do have a problem with it when it’s used to mass produce images for profit.
- I also think it is essential that people be transparent about the use of AI in their work, especially in commercial contexts.
- I don’t think AI is going to be The End for artists; there will always be a need and/or desire for original work.
The one thing that HAS changed since then is that I got my wish about stock content sites being more discerning about how they serve AI-generated content. Now if only I could win the lotto too! XD
Short and Sweet WTF 
Leave a comment