I signed up as a beta user for an AI art generator that uses keywords (text) to create images. The first image I asked it to make was one using the keywords
korean queer mermaid liminal space and these are the four options it generated:
It’s pretty amazing that a computer made these images, and that is obviously a marvel and an incredible feat of technology, although if you look closely you can see where it still needs a little bit of work (for example, in the bottom right panel the face is fairly well done, but I’m not sure what’s happening with the body or why the tail is detached, and the bottom left very much seems like the generator phoned it in).
At this point I’ve tried multiple sets of keywords with varying results and am still playing around with it (there are many many weird and kind of unintentionally creepy ones where the computer misunderstood what I was trying to convey that I won’t share here). Like with any new technology, I am curious what implications this has for humanity… it seems fairly innocent and cool to be able to generate art and images in this way, and that could be used in really cool ways (e.g. to increase visiblity and awareness of queer Korean mermaids), but unchecked, it could also be used for nefarious purposes (the discourse is probably similar to that around deepfake technology). My mind often goes immediately to the worst case scenario in general, which is something I am trying to manage.
In the meantime, here is one of my favorite and hopefully not evil sets so far, using keywords
cassette tape moon 80s landscape colorful --w 3840 --h 2160 (I guess it got confused by “cassette tape,” but otherwise I think it did a good job of capturing the vibe I was going for):
I liked this option a lot, so I asked the generator to riff off of it and make some variations, and this is what it came up with:
Am I an artist now??