In my written piece, In Pursuit of Skynet, I posited that the more common use for "A.I." would be to democratise art in the present and perhaps in the near future spark off another Industrial Revolution. While the latter remains to be seen, the former is already a reality, with some caveats.
I’ve always gone for a very particular visual presentation with my written pieces. Often I pair heavy use of metaphors with visual representations of these metaphors. As an example, if I use creatures from Greek mythology as an analogy of current affairs, I would use a picture of a minotaur or a coloseum. These are often easy to find as many artists have imitated classic art depicting similar.
It becomes more of a challenge when I have a very specific metaphor/visual analogy I’m going for and need an image that lines up with what I’ve pictured in my mind. Often, I would go for what is closest to what I had in mind. However, at times I can’t find anything remotely close to what I'm looking for and thus have to modify my metaphor to suit the image I’d like to use. As an alternative, I could hire a digital artist to render the image I would like to my specifications or learn the skill myself. This would cost time and money. While I do value my craft, I have no interest in spending copious amounts of money on random musings I post on the internet.
Enter A.I. Image Generation.
For my latest written piece, American Godzilla, I wanted to have a split face of President Trump and Godzilla (2014 rendition) mashed together. I also wanted to have extra images of Godzilla to use in other places in the article. I searched for royalty free imagery in my usual hunting grounds and couldn’t pick up anything. Additionally all the images I found that were quasi-relevant were all "A.I." generated. As Grok’s image generator is free to use at the moment I opted to use it to generate the imagery I needed. A few prompts later and on my third attempt I was able to get the imagery I wanted. No skill required from me. In addition, I was able to have the images generated in the particular art style I like them in. Often A.I. generated images look like A.I. generated images. They have a particular "sheen" to them that is noticeable and one I don’t really like. And yes, all work and no play will not make Keli a dull boy.
While A.I. is supposed to have been trained on publicly available information or at least information it has been granted access to, such as user data on Twitter/X, the images it generates may not necessarily be in the public domain. I created an image of Godzilla using Grok 2 but I suspect that the likeness of Godzilla is copyrighted by the company that created and owns the IP. This may be why it was hard for me to pick up royalty free images to use that were not A.I. generated. Additionally, while the images were accurate in the depiction of what I needed, I still have to work on the images generated to have them in the proper format and elliminate some unwanted details.
