No, it isn’t, at least not as a hard rule. Soundalikes are absolutely a thing and are very common. They can be found to have violated an artist’s copyright, but it’s absolutely not universal.
It has no idea what hands are, no model of how bodies or anything in reality works. That it can’t count the fingers is secondary to it not knowing what fingers are. Or what counting is! Prompteurs get frustrated when they try to specify things like “THE MAN HAS ONLY FIVE FINGERS” and it makes no difference.
I suppose it’s all coming, mixing text-to-image with scene analysis and getting the two talking.
The artist seems to have not realized that it would be T. Herman Zweibel sitting on the beach generating look-alike art, not him! He should be off at the side, shaking his fist and wearing a barrel!
Yes, it’s tiring. I won’t pretend I’m an expert, though.
The rest of this isn’t aimed at you specifically, Mindy. I’m not trying to drag you into this… Just enumerating my “sources” such as they are
I’m basically synthesizing a few things together:
- That Google Images and Google Search exist and are definitely commercial. And that Google “won” Authors Guild, Inc. v. Google, Inc..
- That this guy in response to a C&D letter over him extracting and compiling the stat blocks from D&D manuals didn’t get sued by WotC when he responded to the C&D by expanding the project to cover more manuals and posting why he had every right to extract and publish what he considered uncopyrightable expressions in the manuals. …Which I’m taking as support for his explanation of what isn’t copyrightable. (And that you should contact your own IP lawyer if you need advice on your own situation)
- That allegedly the lawsuit’s description of the technology is inaccurate and misleading to the benefit of the plaintiffs. And that they sued the “open” Stability AI and Midjourney, but not the “closed” OpenAI (DALL-E and ChatGPT)
- Gut feelings!
Obviously, IANAL and not cut out to be one… I couldn’t believe it when the Java APIs were determined to be protected by copyright
I’m just saying, while debating the legality of all this, maybe there should also be a more meta conversation about all this, too - like the whole ball of wax with copyright regimes, IP, AI, etc as it already exists and whether it should exist as it does… less about what is, and more about why what is IS, and if we can actually make work our way towards something better.
Oh, like spend more time on discussing why we think things are OK or how we should change the future to be better, less on straining ourselves to be experts just to try and predict the future? …If so, that is a reasonable suggestion.
That seems to me that this engagement with AI (or the crypto shit, etc) are raising some interesting questions that should be probed further. If they are a challenge to our current mode of cultural production/reproduction, shouldn’t we be thinking of that, too?
This is completely wrong in my opinion. The only people who will benefit from the advanced “AI” algorithms will be billionaires, not artists. It’s just another tool to concentrate power and wealth in hands of the few.
To be fair Onion political cartoons are often cynical and intentionally off base
I hope this isn’t just stoking a fire, but if someone could get a trained model to spit out a recreation of a specific work that was visually similar enough to trigger automated takedowns say- could you then say that ai art is stealing, or could you only go after the person who wrote the prompt that reproduced the work? (or the person who trained the model?)
Yes. Like, I want these tools to exist for people generate images for their own personal use. But I definitely don’t want to see companies using it to flood markets with junk, or even for individuals to be flooding art hosts with a dozen generated images at a time all on the same theme. I don’t mind, say, someone spending hours tweaking a single image and uploading it, but just uploading en masse all the ones that look “good enough” is disrespectful to literally everyone else.
And as always, I want some sort of UBI so people only have to work for self-respect, recognition, and luxury… At least, until I remember how unsustainable our society is and how we don’t need to redistribute our wealth but outright liquidate it… and stick my head in the sand again.
I hope so, but it’s hard to tell nowadays. A lot of AI “art” art enthusiats seem to genuinely hold that kind of belief. On the other hand I remember The Onion article about cryptobro moving from crypto to AI, and proclaiming it’s the future after generating a few images.
My gut feeling is that it’s not much different from transcribing a book using a word processor or scanning it w/ OCR software and redistributing it, or the same with a digital painting program and a piece of art. “AI” is just software, it’s not a person (or other legal entity). Whether it’s an accidental infringement or intentional (since ProseAndKahn brought up the inverse ratio rule…). As far as the software goes, I think the question is whether the program is primarily for facilitating copyright infringement.
Arguably it does facilitate infringement a bit more easily than a word processor or digital painting software, since it will try to create something that isn’t nonsense which means it’s more likely to produce an infringing work by chance compared to mashing the keyboard or squiggling with the mouse… But it’s still probably going to spit out something unique, which can not be said for scanning software (hopefully).
Recent court rulings have been expanding on that definition. And we may not agree with it, but it absolutely opens the door to “this art looks a lot like mine, and hey look, they used my art without my permission or compensation to set up their algorithm!”
I should be more clear: This thread makes it obvious that a lot of new (or new-seeming) issues need to be worked out: ethical and technical issues. My comment has nothing to do with “How AI works”… It’s just meant to re-introduce some of our current working ideas about how WE work; do WE want to participate in an experiment we’ve never been informed about. The lawsuit is for the people who create. If I, as an artist (photog, composer, designer) have a work I don’t want drawn into this speculative startup technology, I should have something to say about that. That’s all. We collectively don’t all know how AI works; this lawsuit will help clarify (and Butterick is very well prepared to make a solid argument), but it requires that everyone potentially involved share their POV. Because like with FB, Twitter, IG and a bunch of other recent tech, we are all at risk of being exploited without our knowledge or consent by actors and algorithms whose interests haven’t yet been revealed or even imagined.
But isn’t that what REALLY matters? The intricacies of new tech! Shouldn’t we all blindly worship new technology, rather than figure out what it’s doing, how it fits in our existing culture, and what purpose it actually serves? /s
Just let the AI flow through you! /s
Hmmmm… maybe.
Zoom out: Include all the artists arguing here and elsewhere that they don’t like the tech, because… they don’t understand it. Both are happening at once. This thread shows that we’re all still working it out.
Legally, definitely not ‘the AI’.
That’s not a person or even a legal entity of any kind. It’s just a computer program.
The law there might change - there is a chap who keeps bringing claims trying to claim intellectual property rights for (as in on behalf of) “AI” programs. So far he keeps losing and I doubt that’s going to change anytime soon
The programmer who wrote the AI? Maybe. The person who input the prompt? Maybe.
The person who takes the produced image and publishes it on Instagram? Almost certainly.
Although it would need to be a lot more “similar” than just triggering automated takedowns.