BUT IS IT ART?

[gwpriester]

I read an article on AI in a publication from my Alma Mater, Art Center College of Design, that compared AI generated art to the introduction of the computer for design in the late 80s.

I was in the small group of advertising art directors and graphic designers that saw how a computer could be a valuable tool for the work we did. And it made me think. Hmm? Maybe I need to get behind the ball on this AI generated stuff.

And I can see many examples of how I can use AI to generate patterns and maybe try to teach it how to make depth images (which would be major).

My first AI generated image was the result of a simple instruction: Imagine a frog on a skateboard going down a steep waterfall. And in moments I had an image.

There are several Facebook pages now that are crammed with AI generated images, some of them incredible, but is it art?

Most of the images seem to look alike. As if as an art director I gave an assignment to an illustrator to illustrate a specific concept.

And, sad to say, AI generated art is going to replace illustrators and photographers and graphic designers.

How do you feel about this? Discuss.

[handrawn]

first define what the art is for

if it is to sell m**donalds burgers, et al, it don't matter a damn to me

if it is to make people think, then there has to be more to it because AI art is [ by definition] derivative not original - it takes on the mantle of the hack artist of old, albeit faster and with a much bigger 'portfolio'

philosophically speaking if all that is ever generated is AI, then we are in Jonathan Swift territory where everyone makes a living by taking in everyone else washing....

hopefully it will be used as a tool - like a reference drawing, to boost creativity, but I do not hold my breath...

[gwpriester]

Generated with Bing AI

[Chris M]

I've been looking at AI for quite a while, and (from a personal view) I find specific tools and ChatGPT very helpful, but not so much any or the 'creative' AIs. Both art and music generators will churn something out very quickly, but it's rarely even close to what I was aiming at.

One example of how I set out to use AI vs how I ended up using AI.
Objective: create simple, preferably vector with no shading, images for T-Shirts.
Result: Hit or miss if I got anywhere near what I described. The short time it took to create the images adds up to be mind-numbing. I found it quicker, and much easier, to scroll through Leonardo's gallery, or hit up Lexica and search existing offerings.

A lot of generated graphics are relatively small, 500 to 1000px and I want them at 3000px on the longest edge. If I have a piece that has no shading, or one I think would be okay if it turned into block shading, I totter off to Vectorizer and create an SVG. This I drop into Xara and add frames, colour changes...anything I can do quickly and easily. I aim to end up with 4 or 5 variants of the same image.

If the image has complex shading, I toss it into an upscaler. I use Topaz for this as I have it installed, but there are online ones too.

I sometimes remove backgrounds with varying degrees of success. Again, I have installed software that does this, but there's online AI options.

So, my plan was to have AI: create vector images from my prompts - create SVGs from said images, and, possibly, do a bit of tweaking in Xara.
What I ended up doing: searching existing work and throwing it through a series of AI software to end up with something I wasn't envisaging in the first place.

[gwpriester]

Lovely image Chris. I think the interface that I have seen with Midjourney needs a lot of help. I've only started exploring so I have so much catching up to do. But the ability to create an image, then refine the image with additional input would really be helpful.

I use Topaz Gigapixel AI and Sharpen AI a lot. Even with some of my stereogram patterns which never seem to be as sharp when I export from Xara as I would prefer.

[Chris M]

Lovely image Chris. I think the interface that I have seen with Midjourney needs a lot of help. I've only started exploring so I have so much catching up to do. But the ability to create an image, then refine the image with additional input would really be helpful.

I use Topaz Gigapixel AI and Sharpen AI a lot. Even with some of my stereogram patterns which never seem to be as sharp when I export from Xara as I would prefer.

Home | Leonardo.Ai is similar in content to Midjourney, has a large gallery and shows the prompts used to create the individual pieces and which fine tune model was used.

Free AI Art Generator - Images.AI has an impressive amount of styles (Recipes), but isn't as sophisticated as Leonardo or Midjourney.

BoredHumans.com - Free AI Tools has over 100 different AI tools to play with. All are pretty much rubbish, but it shows what's available with a bit of spit and polish. Worth a look, and you never know, you might be inspired by something.