Database of artists used to train Midjourney AI garners criticism

That said, it seems like a moot point to me. The practical uses of generative AI are not going to be one-and-done prompt-to-image tools. When AI is used more like a brush, the brush strokes the human chooses will still be granted copyright.

Their guidance would not apply for someone using AI as a tool, but said copyright would be thinner than if you'd drawn everything by hand. Specifically, you don't own any of the things the AI "just came up with". If you just wrote a prompt and grinded out some results, you probably own nothing[0]. If you use shittons of inpainting to control, say, the overall composition, but the AI filled in pixels somewhere, then you probably still own the overall image, but that's only because I'm not sure how you'd separate the two in a way that would let you copy just the AI-generated portion. Or, in the case of the comic book I mentioned earlier, they own the text, characters, and plot of the comic book, but not the artwork.

[0] Yes, you could probably just lie to the Copyright Office. Make sure to never reveal your use of AI to anyone, because there's loads of angry artists who would love to tattle on you.

With art, it's more subtle, because there there's no single reference point that lets us determine if an artwork is "true". There are the glaring errors that everyone can agree on - notoriously, human hands - but those cases are improving rapidly.

Human artists might have to become used to tracking provenance. If you work with traditional media, that's easy: Here's the painting. For digital artists, software can publish encrypted, timestamped brushstroke-level histories of the work if we need that level of proof.

People have been trumpeting this since day one of Stable Diffusion releasing, but I'm seeing the same output quality as that day and I've been keeping up.

Keep in mind that in the not so far future, producing art will be as cheap as consuming it, this means that the original benefits society got in return for copyright no longer applies, so why should they protect it?

I love when non-artists talk about art.

If it's digital media, the software can keep an encrypted record at the brushstroke level that can be played back to produce a bit-perfect reproduction. Maybe even write it to a public ledger.

Generally the party asserting ownership has the burden of proof. The standard is "preponderance of the evidence", which generally is understood to mean "more likely than not" or "> 50%". So basically it means if you can prove to judge or jury that there's a >50% chance you own the work, it's good enough.

Also note that in many cases where there's a dispute over the evidence, witnesses are summoned to testify. So you might not have a "full notary log" of how it was produced or all the "intermediate artifacts", but as long as the artist is able to convincingly explain how the work was created, and the other party's lawyers are not able to poke holes in their story during cross evidence, that's usually enough.

Which is, basically, what happens today, if the authorship or ownership of a work is disputed.

That said, I'm not sure whether "assume work AI (thus uncopyrightable) unless proven otherwise" should be the default for other reasons. For one, most quality "AI art" needs some manual adjustments or touch ups, and arguably the prompt and hyperparameters may be sufficient creativity element. I mean, that's basically how copyrights dealt with photography (the mere fact you decided when and where to point the camera with what settings is sufficient for copyright to subsist in a photo).

Where this is more extreme here is that there is no human labor involved there is no invention. On the other hand LLMs make this extra tricky because in one way output they create is objectively unique but subjectively/culturally it's not.

To answer your question if you take AI generated image and change it enough for it to stand as it's own unique thing you could for sure claim it as your work and it would be eligible for copyright.

in the same way you can brute force an MD5 Hash if you had a few centuries, I guess. I don't think "monkey's making Shakespeare" is a good metric for when to determine if a piece of art is unique enough to be scraped.

Tracing is looked down upon in the art community and this feels way too close to tracing if they just have these images in a database ready to reference. You can't store copyrighted movies nor music on such private databases (one of the few times I will ever utter the words "thank you DMCA"), I don't see why art pieces would be exempt.

In this context, I don't see a principled difference between the model weights and really good compression. If I send you a gzipped copy of the latest bestseller book it's still copyright infringement. And it would still be infringement if I shipped it inside a software program that can _also_ reshuffle the words in a bajillion different ways, if there's a "copy" of the original work in there.

I don't think fair use saves us either, at least not the model authors, because the whole selling point of these models is to replace artists. Yes, artists can use them as tools, but that is far less lucrative. The valuations and hype being thrown around specifically come from, among other things, being able to cut the creative class out of their own business. No court is going to look at that and say "ok, yeah, sure, it's perfectly fine for you to be using other people's work to train on".

An individual using an AI art generator may still wind up getting novel output that isn't obviously infringing to anything in the dataset. If that's the case then they probably haven't infringed copyright. Or at least, it'd be difficult to make a case around it.

You appear to be refuting a slightly different point, though. When OpenAI was making Dall-E 2, they found that duplicates in the training set would incentivize memorization of specific images, as you said. This is more like finding a secret cheat code that would make the model regurgitate everything it had remembered, regardless of how "incentivized" it is to do so or even if it had been aligned to not do that.

My personal argument is this: the primary metric that the training process attempts to minimize is perplexity. This is how good the model is at guessing the training set data. Base models are specifically being designed to compress huge amounts of text and we just so happen to accidentally get a decent word calculator out of it. The alignment fine-tuning that happens later adjusts the model to prefer answering questions, but the underlying memorized data is still there.

>The way these algorithms work is public knowledge, there's not really any black boxes in the hands of OpenAI that would be relevant here.

Nope. Modern GPT is entirely a black box. OpenAI stopped publishing model weights the moment Elon Musk stopped writing the checks. Hell, they don't even publish the model architecture anymore. How GPT-4 works, even on a basic "this is how many transformer layers and attention heads we're using" basis, is a trade secret.

Even if you have model weights, the actual meaning of the learned model parameters has never been known; there is active research on figuring them out. One particular problem is polysemanticity. If you look inside a particular hidden layer, you won't see a single "dog" or "cat" neuron in its hidden layers. You'll have a 512-dimension concept bouillabaisse with "dog", "cat", "bird", "guinea pig", "kangaroo", and so on all floating around whatever shape made sense at training time (even if it implies absurdities like "desk is the opposite of loin cloth"). To untangle this, you have to train another AI to pick out monosemantic clusters of neurons that can then be inspected, and that requires extreme amounts of GPU resources.

Random means there can be copyrighted content in it. It shouldn't be "random". You can't do such things when making your own piece of media, so I don't see why it's okay when it's a bot scraping massive pieces of the internet for use in some for-profit venture. Scraping was already this gray area in the internet and practices like these argue against such techniques.

>Then, when a model is trained on that dataset and that model is used to make other outputs, are those outputs actually derivative with a direct link to some origin?

That's the million dollar question (literally. Or more like 10/100 million). I imagine like existing copyright it depends on how far out you derive. Arguments above that suggest the ability to regurgitate the exact or very close to piece of data would certainly lean towards "yes". In which case it implies they have that piece of data in their database, instead of just being a "reference".

That feels too close hosting off of copyrighted material in my eyes, but I don't have the full picture of how and what these LLMs contain.

>If I generate a landscape using an AI, to whom exactly would I even owe money?

Like any media creator/producer would tell you: depends on thr license. For example, I Google "mountain" and my first result is this :

https://unsplash.com/photos/aerial-photography-of-mountain-r...

License says "free to use under the unsplash license", which is surprisingly lax. But there are two stipulations

    Photos cannot be sold without significant modification.
    Compiling photos from Unsplash to replicate a similar or competing service. 

And that's one of the more generous results. Another result came from CNN but sourced from Shuttershock. Here's their license :

https://www.shutterstock.com/license

I'm not going to read nor summarize the license, but I'm sure CNN at the bare minimum needed to ask permission and likely negotiate to use it in their news post.

So yeah, an "ethical" training model will be doing this for every single image they use, if the aren't creating nor taking the pictures themselves.

>Equally to all contributors in a dataset?

Don't know. That would be for the courts to decide. If it's anything like streaming music and movies, you would be compensated proportionally to the amount of "usages" your work gets when generating pieces. Doesn't necessarily have to be equal, but it may surprisingly equalize out since algorithms are at the helm and not brands trying to stand out.

It's certainly be it's own rabbit hole to explore though. If we ever get that far.