As concerning as the implications of artists (as we know them) becoming obsolete may be, the machine learning algorithms behind the newest trend in computer-generated imagery are still really dumb to the point where referring to it as "AI" may as well mean "Artificial Idiot". These are tools that still need data input in order to be able to spit out the pieces that they make, and barring any actual glimmer of intelligence that's not part of a Skinner Box experiment it's highly unlikely that they have any understanding of what they're making and why they're doing it.

Rest assured however that while artists in the traditional sense may be on the losing end, there's still a venue for quality control and troubleshooting of these programs that would be a natural evolution.

My view on AI generated art is this:
At first, it's a neat little gimmick. Feed an algorithm a phrase or sentence, hit enter multiple times, then see what pops out. It was fun as a neat little gimmick, and could even be useful for creating references for people to give them an idea of what to draw.

But when people start using it as actual art(Like that one competition where someone submitted AI Generated art and won)...it just turns it into something hollow. Art should have a distinct style to it you know? Yes, some do copy other styles, in the beginning, but not all the time. But the AI Art being posted always has the same smooth almost corporate like style to their images, just becomes uncanny you know?

AI Art is Hollow, for you know a machine made it by using other images as a reference, essentially smashing images together to make a new one. Turns what should be something fun and gimmicky into nothing but a shallow, empty, "easy way" to make art. And what really grinds my gears is that there will be people who will pretend they made it when it was a bunch of code that made it.

To put it shortly: When used as a fun little gimmick or a bit of a reference for inspiration, AI Art shines. But when its used as actual art...it just turns it hollow and shallow.

Hollow is a good way to put it. It sorta reminds me of the practice of asset flipping in the indie games industry. only a robot is doing it.

Like how fast food hasn't destroyed conventional restaurants and private businesses. Because the more automated it becomes, the more trash it becomes. Amazon was once thought to be the great destroyer, but now its just an ebay for Chinese crap where you have no idea who you are actually buying from. And you better believe those future shitty Disney AI generated movies are gonna be marketed at regular ticket price despite being made for a fraction of the cost.

And are people gonna be typing "Horror thriller based in Dairy Maine" into a box and getting a Steven king novel that's at least half as good as the actual thing? and if they do it multiple times will it be 100% new with zero repetition?

Those students who Gibli's Miyazaki told off after showing him their ai generated walk cycle might think they have the last laugh, but their work is still ugly as hell.

This is something I've thought deeply about and researched for quite a while now.

The whole thing brings up many fascinating questions, as deep as the very question of what is art, and what really is the difference between programming a computer with a library of images and some language and what any artist does with what they've managed to absorb from history and culture. These are things well worth pondering.

My conclusion at this point is that AI art is primarily and best used as a tool, and can be good for that. It can be entertaining and surprising, and that's good. Like any tool, it can be used for good or ill. It's pointless to fight it, the cat's out of the bag.

The new blanket anti AI policy here is particularly short-sighted for instance, among other negative things. They would do much better to just have a strict policy of tagging it honestly, because by making it "Illegal" they'll just drive it underground and people will lie about it and use it anyway, and they'll be hard pressed to prove it. Historically, making art illegal has never proved to be successful, or even possible in the long run. Look at sampling, as a prime example; once technology makes something possible, it's going to be done whether the "Authorities" like it or not. These technologies, being digital, quickly become ubiquitous and are impossible to stop. Furthermore, they're deeply democratizing, and make amazing recombinations of existing material accessible to people who would never have had access to them before. Whole new artforms become possible or even inevitable with new technologies. This is a good thing. It's pointless to fight a tsunami, you have to go with it.

AI art doesn't compare with human generated art, and it never will. It's just a different thing. No doubt for some purposes cheap or lazy capitalists will exploit it for their benefit, such as for cheap advertising art, but they're still going to need an artist to generate the art and edit it, for instance.

This stuff is still in its infancy, and really misunderstood, and, like most things that people don't understand, they fear it. I don't happen to think it's something to be feared, it's something to use and play with as a tool or for amusement. There are occasionally some pretty impressive pieces, but those are cherry picked from a sea of broken and weird results. Personally, I find the more "Broken" results fascinating, particularly as documents of this form in its present infancy. Even learning how to prompt these programs and combing the results for "Successes" takes a certain human touch.

If and when we ever have a true sentient AI generating art, it would also be a completely different thing, and bring up a whole new set of fascinating questions.

*Mann vs Machine theme*