AI/LLMs AI art bans are going to ruin small 3rd party creators

[Morrus]

It's sad how legit this technique is.

It's also a common marketing strategy. Make a post or ad with a deliberate error in it and the engagement multiplies tenfold as the world gleefully corrects you (sharing it widely in the process).

 

[RealAlHazred]

It's sad how legit this technique is.

"We'll be proving a lot of stuck up buffoons very, very wrong. Which, trust me, is the very best thing about science."
-- Deborah MacGuiness, paleontologist (1899)

This sentiment is also supported by the Internet in general, not just science.

 

[Bill Zebub]

Well that's the problem isn't it? If you get so used to using LLMs to get information, you start to lose the knowledge that we all used to have of where to go for good information. We used to know this stuff. It's like a skill that's atrophying. You lose the basic ability to look things up because you no longer know how.

Uh....that wasn't at all what I was saying. I put a lot of effort into researching the sources I found long before LLMs were available.

EDIT: I would say, instead, that the change from visiting the library to using the Internet has necessitated checking your sources. And that experience has, in turn, made me realize that I was probably too trusting of sources in libraries.

 

[Morrus]

Uh....that wasn't at all what I was saying.

Um. No? It’s what I was saying.

This also isn’t what you were saying.

 

[Bill Zebub]

The antecedent of my "that" wasn't what you said, it was your apparent misinterpretation of what I was saying. Since you wrote "Well that's the problem..." I assumed the antecedent of your "that" was what I had said.

But maybe you were referring to some new, undefined "that." If so, I failed to follow it. My fault, clearly.

 

[soviet]

And that experience has, in turn, made me realize that I was probably too trusting of sources in libraries.

? All library books are extensively fact-checked.

At least, that's what AI told me.

 

[LuisCarlos17f]

I have looked for more information and we have got the example of the artwork "A Single Piece of American Cheese" but we have got also the cases Zarya of the Dawn (2023) (Graphic novelist Kristina Kashtanova used Midjourney to generate images. USCO protected the text, selection, coordination, and overall arrangement of the images (as a compilation), but not the individual AI-generated images. The human labor involved in the assembly and structure was deemed sufficient to grant partial protection) and the case in China Li vs. Liu (2023)
A work assisted by AI can be protected by copyright. I believe this detail is an important distinction. I suppose professional artists will take care of recording evidence of their own creative contribution.
My opinion is that AI will paradoxically boost creativity because it will be used by amateurs for amateur works who have very creative minds but are too clumsy to put those ideas into practice. But this saturation of creativity ultimately leads to only the freshest and most original ideas remaining in our memory.
I highly doubt AI will be used for overly photorealistic images with human characters due to the risk of the uncanny valley, but we could see many works with a 2D art aesthetic.

And personally, I enjoyed creating and publishing my PC species ideas althought the power level wasn't right.

 

[Umbran]

Like I said, your understanding of history, moral rights, natural rights and common-law rights is seriously lacking.

Mod note:
One wonders why you felt that making this personal was appropriate.

Folks with a dozen warnings on their accounts should take more care than this to avoid moderator attention. This may easily become a discussion about whether your posts are worth the grief and disruption they cause.

Be better to your feĺlow posters, please.

 

[FrogReaver]

It really does, but to concede it would be to concede the point, which you don't seem to be capable of doing.

I think if you could correct me, you'd do so. Instead, you're going with vague hints of how it could be wrong, thus not actually saying anything while trying to dismiss my point.

I think it's more likely that we're not going to get an agreement because one is trying to dance around the issues here rather than engage with them.

Once again, saying things without backing them up. You have no evidence to your claim, so you attempt to demean mine.

But I find this all to be sour grapes. Trying to find some way out of simply recognizing that these laws were for the protection of authors, because without them it would be a harder cost-benefit justification to try and innovate or create.

Perhaps if you read it yourself, you could actually make that distinction that Claude failed to do.

Once again, someone who did not actually read the article.

Perhaps if you weren't so quick to try and ask Claude what it meant, you would actually read the introduction.

The pushback was simply because y'all were looking to discredit the piece without understanding it.

Because y'all didn't even read the intro before you started trying to discredit it.

But I can confidently state I read more than you did because I was concerned with actually engaging with it instead of immediately discrediting it.

You can say "independent conclusion", but seems weird that you and the AI somehow managed to reach the same weird, limited conclusion.

You can say this, but I don't think anyone could read the intro and somehow come to the immediate and wrong conclusion you did. It just comes off as incredibly weird and suspicious, to say the least.

After reviewing this context I am done with this particular conversation.

 

[pawsplay]

How is a ban on something I don't use in my business going to ruin me?
AI is pretty interesting, and LLMs have made a big splash lately, but let's be clear. As it stands now, AI is not an acceptable substitute for human labor, aside from any legal or moral issues that might arise. From a business standpoint, I think the cultural landscape for it is a landmine field, I think it is very risky to invest in a product people are turning against, and which doesn't have hardly any decided case law (in the US, they determined a piece of software cannot be the nominal author of a work, whoop de doo). Generative AI, in its current form, has three potential actual uses:

• As a toy
• As an adjunct to the productivity of someone who already has considerable skills they can use alongside this output
• As a way of conjuring up garbage no one wants.

It's not just that AI has taken over search engines, it's taken them over with junk people could easily generate for themselves. In the sense that a third party publisher could use some prompts to produce some "art" (that is, illustration), curating the results will not be any better or cheaper than hiring an artist. The alternative is to just accept what it gives you, which I will tell you, is not a good way to produce quality products.
So, actually, such bans might ruin some "publishers" but mostly people with little to contribute creatively. It sweeps the field for people willing to use human creativity. Perhaps in some future we can look at improved tools that can reliably produce quality, context-sensitive product, but that's a good way off. LLMs really only solve one problem, how to create the surface or appearance of things humans make. They aren't thinking machines and they don't have taste or integrity, and they are foul at following directions. Getting an AI to generate an image is like working with the stupidest freelancer who ever lived.
I don't agree with strong copyright claims in general, and I don't agree AI training is "theft" and or possibly even infringing. But I also take issue with how a half-baked product has been used to push creativity out of markets and visible communities. There's also a huge environmental cost for people who use it casually. There's nothing unethical about driving a semi truck, but there would be a real problem if we all drove them. And there are ethical issues around using generative AI to impersonate other people or their styles; while you can't "own" a style I think there is conceptually a problem with using software to actually replace a person's body of work, just as there is with stealing their face or using their name without permission.
So, until generative AI can compete with the quality AND sensitivity of human work, there is no reason to use them for things like this. At such point as they can roughly imitate a talented but unimaginative freelancer, we can talk about the moral and economic models we need to sustain our culture. Right now I think the biggest impact has been affecting the demand for stock art, or for OC commissions. That trend can be reversed by simply refocusing on hiring humans and their work. It's not a matter of putting the genie back in the bottle, in this case it is simply a matter of recognizing things of value versus empty imitations.