this post was submitted on 26 Aug 2023
329 points (93.2% liked)
Technology
59243 readers
3428 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
My guess is that AI’s first big victim for graphic design will be stock art. Previously, crap like that background asset would just be stock purchased from Getty or Adobe stock. Now it can be generated.
I’m already starting to use it instead of paying for bullshit licenses.
I've been using AI for school and work, as God intended: give it the raw, have it do the grunt organization work, and then proofread to correct anything.
There is very little to say that hasn't been said. For an example of our limitations as humans, there's only 50ish unique plot lines in the English language. To expect each person to be completely original is asinine.
It's a tool, one of many in my toolbox. People who are just flat against any and all AI or LLMs are behind the curve.
How would the unique plotlines be determined by the language they're told in? Why would the amount of plotlines be based on human cognitive capabilities? None of this makes sense.
Either way, "unique plotline" doesn't mean anything, from the perspective of literary or narrative studies. There's no universal, objective way to dissect narratives, and they cannot be boiled down to a distinct number of basic models. There have been attempts to get to the most fundamental narrative model (Greimas, Campbell), but they're far from widely accepted.
Art is, by itself, not something that has "the curve". If you're doing something with very practical goals and need hyperproduction, sure, but art is not necessarily made or consumed with such a logic.
Pretty much.
People very frequently complain about AI taking the jobs of artists. But if the money was never actually going to be put on the table for artists to claim, I really don't think that was going to help much.
That doesn't mean I hate artists what do, absolutely not. It's just that artists are people and people are limited in how much they can do at any single time.
For the past couple of months. I've currently been waiting on multiple artists to finish up their commission queue. And one of which I'm worried I'll have to turn away because of a variety of life changes in my life that's led me to losing my job and me having reduced income.
As of right now, the costs of generating a picture with a tool like Stable Diffusion or DALL-E has been pretty low, the former even being free if you have the right hardware. And these systems manage to be almost always available, as well as being capable of working in a matter of seconds.
Of course, that doesn't change the fact that these tools are only good at painting the bigger picture. They have a tendency to choke on the smaller details. And I would personally rather wait for an actual person to be available to work on something original that's also capable of filling a niche that AI models have yet to be trained on.
This entirely disregards the fact that the training of these models was done on human artists' work without consent or renumeration. As it is, it is not "AI", It is just a glorified plagiarism machine. Not to say it isn't impressive, but it has already stolen work already done by artists and further stealing upcoming work by mashing together older works.
There's ways to do it ethically by training on artwork with permission kind of like how Adobe is doing it, but that isn't going to have as wide of a reach as the other free ones.
You keep using that word "stolen", I do not think it means what you think it means.
Also, AIs do not "mash together" works from their training sets. This is a very common and very incorrect conception of how they work. They are not collage generators or copy-and-paste machines. They learn concepts from the images they train on, they don't actually remember fragments of those images to later regurgitate in some sort of patched-together Frankenstein's Monster.
You're correct but it's still too early and most people haven't spend enough time with AI to fully understand. Maybe they never will.
Like the classic quote says, it is difficult to get a man to understand something when his salary depends upon his not understanding it.
I just asked Wombo Dream to make the Mona Lisa and it did. Sure, you can tell it's not exactly the real thing, but I don't know how you can say it didn't copy any of the actual Mona Lisa original.
I considered including mention of overfitting in my earlier comment, but since it's such an edge case I felt it would just be an irrelevant digression.
When a particular image has a great many duplicates in the training set - hundreds or even thousands of copies are necessary - then you get the phenomenon of overfitting. In that case you do get this sort of "memorization" of a particular image, because during training you are hitting the neural net over and over with the exact same inputs and really drilling it into them. This is universally considered undesirable, because there's no point to it - why spend thousands of dollars to do something that a copy/paste command could do so much better and more easily? So when image generators are trained the training data goes through a "de-duplication" step intended to try to prevent this sort of thing from happening. Images like the Mona Lisa are so incredibly common that they still slip through the cracks, though.
There's a paper from some months back that commonly comes up when people want to go "aha, generative AI copies its training data!" But in reality this paper shows just how difficult it is to arrange for overfitting to happen. The researchers used an older version of Stable Diffusion whose training set was not well curated and is no longer used due to its poor quality, and even then it took them hundreds of millions of attempts to find just a handful of images from the training set that they could dredge back out of it in recognizable form.
People have also copied art for as long as art has existed. You can buy a copy of the Mona Lisa in the gift shop, or print your own. That's why the market for art has always been hyperfocus3d on 'originals'. But rarely are the artists the ones getting rich off their art, especially now. I hate capitalism as much as anyone but if your motivation for making art is money you're in the wrong business and your art probably isn't that good anyway.
Graphic designers aren’t the first. Automation ended a lot of jobs for decades. Ai is just a form of automation.
The wheel is a form of automation
Yeah but the wheel was discovered before rent
They said AI, not automation.
The comment you are replying to linked AI and automation. It makes total sense in this context.
They will be generating it themselves soon enough. I contributed some stock photos in the past. They recently sent me info about their new contribution pipeline, for content that may not pass the usual quality threshold, but will help train the models. If they do it right, who knows, maybe they can get better results worth paying for.
The fun part here though is they dont have copyright on that art. If any of the "stock AI footage" becomes iconic, its public domain.
Dicey spot for a studio to be in, but it does save some bucks, so they are plowing ahead.
You should consult with a lawyer first. The amount of misinformation circulating on the Internet about how AI art is all public domain is enormous. That recent court case (Thaler v. Perlmutter) that made the rounds just recently, for example, does not say what most people seemed to be eagerly assuming it said.
im also someone who has been misinformed on the AI art copyright status. could you explain how it actually works or link to a resource that does? i tried searching around for a bit but couldn't find a clear consensus on it.
It will be really interesting to see how the case law develops. Personally, I am more interested in things on the IP side. A lot of lawyers I work with currently view LLMs like a shredder in front of a leaf blower. Which, it kind of is.
Neither do they have copyright of the stock art they used to purchase. The complete piece, however, including pip boy, is not AI generated. Someone put this together, put effort into it, which easily qualifies it for copyright protection, even if the background is AI generated instead of bought stock art.
If you're talking about that recent legal case, look again. The artist made the claim that the AI was the sole author, but that he should own the IP. I think the vast majority of people would claim that, in it's current state, the AI is a digital tool an author uses to make art. The recent ruling just reconfirm that A machines aren't people, and B you can't just own another author's work.