There’s one type of AI image that seems particular common – older women with enormous flawless breasts. I’d guess it’s a single person with a very specific kink churning them out, but maybe there’s a whole art collective involved. I would say it’s one for Dr. Freud, but even he might think it’s a little too on the nose.
The image below – with a caption added by Tom Allen – is an example of what I mean. You can find other examples on his twitter here and here.
Tom left a comment on my previous post that made an excellent point: AI imagery is likely to be a great source for caption creators. It avoids the issues of copyright and unwanted repurposing while giving a lot more control to the caption writer. Right now the image prompt writers and caption writers are likely different people. But once the tools get mature and widespread I can image a lot of new content from people writing image prompts for the particular scene they want to then caption.
You can see more captions from Tom on his tumblr and twitter.
Thank you for the shoutout.
I’m noticing a trend that some creators are using AI to put the faces of real or known women into their AI images. There are some apps that allow you to use pictures as starting prompts. I’ve tried a few using Danica Collins, and the several I got ranged from interesting to downright horrifying. So yes, there are still limitations in the algorithms.
There is also an “uncanny valley” effect with some artwork. It’s the kind of thing that looks okay on your small phone screen, but when you see it on your 20″ monitor can be unsettling.
I have a handful of pictures for which there are very “mature” looking women with the body of a 30 year old. I don’t think that’s an issue with the creator so much as a limitation of the software; some of that will work itself out as the creators refine their skills in using AI as an artistic tool.
BTW, I mentioned in the comment in the previous post that I have asked permission from some of the models to use their images. I’m now asking the AI image creators (when I can find them) if I can use their creations to caption.
An interesting dilemma will occur when I want to use an image that has the face of a known model on an AI generated body. Who “owns” say, an image that has the face of the above mentioned Danica that has been merged into a creation like the one you just featured?
Interesting times ahead!
Love to understand the prompts you’re using to achieve optimal results please?