Be realistic and practical here, in what situation are you going to start typing into a computer "A teddy bear walking in a park" and then start to do anything with the images it generates in a case where you would've entrusted and artist or designer to make something for you ?
What about when you need different scales, different colors, different this and that, are you going to sit there and keep prompting for this ?
I've worked with people who are really good with Illustrator, they can change and produce images for different purposes in a ridiculously short amount of time already. The ting is, even if I could recruit AI to do it, I probably wouldn't and just want someone else to take care of it. To make favicons and screw around with thumbnails and vector images and the list goes on.
There's also the fact that, some people have an eye for style and application of it. Even when I've tried to use DALL-E, which I never have successfully because it always generates something weird enough to be unusable, I've not been confident that my own personal selection is good enough, I'm not a designer, I don't have an eye for art or color, or style etc. It's not my profession to know what's nice. I don't want that responsibility, so I outsource it.
The problem arises if a bunch of studios start adopting AI assisted workflows to increase productivity and the studios that are holdouts get left behind because it would no longer be economically viable to continue animating without AI. It could be similar to how traditional cel animation has declined in use. If AI assisted output clears the bar for consumers, then it's probably good enough for managers.
People using AI professionally probably won't put the raw output of DALL-E straight into the finished work either, I've seen artists that use Stable Diffusion to generate a base image and then do heavy amounts of editing with Photoshop or similar.
We're talking about digital art and it would actually piss me off (a lot) if I didn't have original copies of the art I paid for to be modifiable later. Especially if I was using the work commercially.
I can of course imagine a situation where I might be able to get a PSD file or vector from DALL-E, but honestly I can't imagine myself sitting there telling DALL-E how I want it to modify my PSD in minute details, I'd just pay someone to do it.
This is where I think there is a practical misunderstanding when people talk about this subject, there is a lot of fiddly work and minutia that goes on with art which I just think people don't realize.
What about when you need different scales, different colors, different this and that, are you going to sit there and keep prompting for this ?
I've worked with people who are really good with Illustrator, they can change and produce images for different purposes in a ridiculously short amount of time already. The ting is, even if I could recruit AI to do it, I probably wouldn't and just want someone else to take care of it. To make favicons and screw around with thumbnails and vector images and the list goes on.
There's also the fact that, some people have an eye for style and application of it. Even when I've tried to use DALL-E, which I never have successfully because it always generates something weird enough to be unusable, I've not been confident that my own personal selection is good enough, I'm not a designer, I don't have an eye for art or color, or style etc. It's not my profession to know what's nice. I don't want that responsibility, so I outsource it.