Home Artificial Intelligence Why Does All AI Art Look Like That?

Why Does All AI Art Look Like That?

Why Does All AI Art Look Like That?

Way back within the Nineties, I went to Art School. It was difficult, fun, and informative and amongst most of the memories I actually have from the time, one stands out with regard to the continuing discussion about AI-generated art.

Life Drawing was considered one of the primary classes all first-semester students needed to take. That is the category where you spend 2 hours a day meticulously studying a bowl of fruit or box and pole, or a unadorned person in a curious pose. These life-drawing models can be arrange and lit by harsh lamps to solid dramatic shadows, so we could observe and find out how light and atmosphere effects surface texture and creates the composition. It’s a difficult class since it is A) extremely technical and B) extremely boring.

It didn’t matter what your interest or major was — painting, sculpture, photography, interior design, Fashion (or, like me), sequential art — nobody likes drawing sweaty naked dudes sitting on a box for 2 hours. But it’s important because that is the way you learn the basics of all art, no matter medium.

One student specifically, had trouble in life drawing. He was, like me, concerned with sequential art. He desired to draw comics for Marvel and didn’t really get why he needed to spend 4 hours every week drawing the sweaty naked dude sitting on a box as a substitute.

The primary week of sophistication, the trainer had us bring a bit of our art. Something we were happy with, to indicate the category. It was each an icebreaker and a method to introduce the category to the concept of public critique.

My fellow comic book artist showed us all a page from his comic book portfolio that he was very happy with, and which he had very obviously traced from a number of published comics. The figures were stiff, familiar, but different, as he’d slathered on his own details and altered barely enough of the figure to make them “original characters” while still leaving enough detail underneath for those of us aware of the recent Marvel and DC publication schedule to discover which issue and page he’d cribbed from. The backgrounds were minimal to non-existent, and the angle was likewise barely there.

Honestly it wsn’t even this good.

I felt embarrassed on his behalf. Not because this childish attempt at art was so excessive and unpolished but since it was so familiar. That’s how lots of us who were concerned with comics began. We’d trace our favourite panels and draw our favourite characters, doodle battles, try our hand at layouts; fail, and check out again.

And this is the reason we were there at art school, to learn find out how to do that. Our professor very patiently and kindly ripped this student a latest one. He was firm but nice about it, but clearly, it left an impression on my fellow aspiring comic book artist. This was probably the primary time anyone had told him his art sucked. And despite the fact that they did it with aplomb and within the gentlest of how, it’s never fun to listen to that what you thought was good work, was only a derivative mess.

Pondering back on it now, that fellow classmate then would have loved AI-generated art, since it did exactly what he wanted: it childishly mimicked the true thing in his head. Actually, a number of contemporary AI “artists” have never experienced any level of critique, until they posted their “work” on social media.

By now you’ve seen some selection examples of AI art (or possibly even created some — no judgment, this can be a protected space). Also by now, you’ve probably asked yourself, “why does all of it look *that* way?”

Chances are you’ll not have the ability to discover the particulars of what *that* is, but you’ve seen enough of it to know at a look at some thumbnail on Twitter or Facebook to go, “right, That was made using Midjourney.”

Ooh, spaghetti waterfall, yum!

Since this can be a topic of dialogue, one I even have some knowledge of, I believed I’d take a moment to stipulate the 6 ways to discover AI-generated art (and why they appear like… that):

. Hair, skin, necklace, clothing — all of it looks prefer it was created from injection-molded plastic. Human artists spend years learning techniques to render different materials so that they have various textures. The algorithm behind AI-generated art uh, doesn’t do texture thoroughly. In any respect. It’s at all times too smooth, or too busy, with no sense or actual thought behind where the small print go. There’s no foreshortening or blurring, no focus. The Algorithm has no idea where the attention of the viewer must be drawn and so just moves it in all places, suddenly.

Statistically, a human has between 4 and 11 fingers. So, good job, Mr. Roboto!

. AI famously doesn’t know the way many fingers a human has, but there are also other cues when you’ll be able to’t see hands. It’s a bit of harder to inform with a bit that’s mimicking stylized cartoonish anatomy (anime, for instance) however the misproportioned eyes and lopsided body construction are a dead giveaway. Also three rows of teeth (why is it at all times three rows?) AI art doesn’t construct on an armature of sketched-out anatomical constructing blocks checked for composition and relationship, to make sure the figure is dynamic but based on some anatomical model resembling a human (or anything). There’s no underpainting. No trial and error. That takes time. And speed is the essence of AI art.

It’s just heads all the way down.

. AI art lacks pers[ective (and I don’t just mean the flexibility to step back and interact in critical self-analysis, though that’s a problem among the many proponents of AI art). What I mean by lack of perspective is, AI art has no concept of fore-middle-background, so the whole lot looks flat. There’s no dimensionality, no atmosphere, no foreshortening. All the things is rendered in soft detail, such as you’re probably not focusing your eyes, just glancing at shapes in space.

How far is to cactus mountain?

. This one is a bit trickier to note when you’re a novice or someone not versed in art, especially cinematography, but when you do see it, it’s all you’ll be able to see: AI images are at all times eye-level head-on, what’s referred to in TV and movie making as American plan. It’s composed as when you the viewer are sitting within the space with the topic, it from arm’s length, with the main target at eye level. AI-generated art can’t do birds-eye view, worms-eye view, 3/4 high/low, Dutch angles — nada.

Saint Don Knots, patron of nipping it in the bud.

. Where is the sunshine source on this portrait? Who the hell knows, because there’s not enough contrast within the shading to inform. Even when there’s heavy shadow/light, AI images have puddles of sunshine from random sources. The Algorithm is incapable of considering dimensionality, and where a light-weight source could be coming from, especially if that light source isn’t within the frame of the particular piece. All the things looks prefer it’s shot on an overcast day, or under soft, studio fluorescent light. Or worse, it has a nasty case of vaseline lens, and the whole lot looks haloed in diffuse light from nowhere.

hare it when my hair turns into a necklace.

. Hair that blends right into a necklace; superfluous limbs; background figures that emerge from nowhere: these are only a number of the classic tells of AI art, where the small print just kinda get smooshed together. The rationale is, AI can’t handle tangencies (when different objects line up visually so they appear like they share a surface). So it does what any inexperienced artist does and fudges it. Obscures it in shadow or smudges over it with greeble (that noodly stuff that appears like an object covered in spaghetti or melted wax).

Proponents of AI art have argued that it’s still the early days, they usually’ll get these limitations sorted out. Perhaps they are going to, possibly they won’t. But someone should tell the eager beavers who’re applying for jobs to interchange half a dozen concept artists to slow their roll and possibly put a bit of more time into learning the basics of why smears of pigment and graphite and chalk on paper or board can persuade us it’s an image of an individual or a mountain or a sunset.

Simply because Bob Ross has given you permission to be generous with the joyful little trees, it doesn’t mean you’ll be able to make a living as a landscape painter.

After 4 years in art school, I learned lots. So did that hapless classmate from Life Drawing back in freshman 12 months. As we were in the identical department, which wasn’t that big on the time, I bumped into him occasionally. During our senior 12 months projects, I went to a viewing of labor by soon-to-graduate students and he had a few pages of comics up. The figures were vibrant, and dynamic, crafted with a skill that he had gained from 4 years of labor and determination. He took our kind professor’s advice to heart and learned find out how to draw a figure in space in order that it had volume, texture, perspective — life.

Perhaps AI Art will get to the purpose at some point where it could make convincing real art. I doubt it, but when it does occur, it should because someone, somewhere took the time to show it find out how to make art, not only trace, copy, and manipulate another person’s.



Please enter your comment!
Please enter your name here