Just a minor post to put up a few things I’ve been playing around with, using the Stable Diffusion AI. It doesn’t do photorealistic images (or at least when it does and they feature humans, it is seriously weird) but with a lot of tries (most not shown here) I got some mildly intriguing artistic effects.
I imagine this is going to be huge, when photo images get better (see Paltego’s latest post at Femdom Resource). Not quite yet, although already I find myself suspecting images of big-eyed ladies with perfect skin on Tumblr of being AI-generated – and some admit to being so. I have always held out against fake images to caption on this blog (photo-shopped actress heads on dominatrix bodies and so on) but eventually, I suppose, a difference that makes no difference is no difference at all.
Anyway, in this post this blog is sticking rigorously to its established policy of never posting anything remotely realistic, so I am not trying to make convincing photos; rather, below you will find some attempts at AI art.
We’ll start with a few ‘artistic’ takes on Gal Gadot (I found that specifying an actress with well-defined features – and her features are beautifully well-defined – led to much better results, as otherwise the AI would produce generic blurred faces).
I played around with a few artistic styles. Most didn’t work well, but it can generate pretty good Aubrey Beardsley (simple line and black-and-white images, plus the originals are frankly pretty kinky already) and I also quite liked what it did when I asked for traditional Japanese dominatrices, although people more familiar than I am with real Japanese art are welcome to disagree!
And finally a few items of ‘steampunk dungeon furniture’. Not sure they’re entirely practical but they are rather forbidding and I for one would not like to test them out…