Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Adobe Firefly: AI Art Generator

adobe.com

Post reply on HN
Print view View original

341–350 of 496 posts

Re: Adobe Firefly: AI Art Generator

#341

Post by roflyear »

pcurve wrote: post #303
user3939382 wrote: post #246

Adobe has lost any good sentiment I had for them with their forced subscriptions and dark patterns. I use their products begrudgingly, it’s a sunk cost fallacy from decades of muscle memory. My reaction to seeing any announcement from them is, yeah whatever Adobe.

Once you get stranglehold of the market place, you can pretty much do whatever you want. Adobe... Microsoft... Apple... Google... all pulling the same lever.

At least with all of those (except Adobe!!) you can cancel your subscription using the same interface you bought the subscription on!

Re: Adobe Firefly: AI Art Generator

#342

Post by astrange »

roughly wrote: post #243
mesh wrote: post #226

Earlier quoted context omitted.

We are working on a compensation model for Stock contributors for the training content, and will have more details by the time we release. The training is based on the licensing agreement for Adobe contributors for Adobe Stock. (I work for Adobe)

I would be very, very interested to see a compensation system that took into account the outputs of the trained model - as in, weights derived from your work are attributable to X% of the output of this system, and therefore you are due Y% of the revenue generated by it. It sounds like Adobe is taking seriously the question of artist compensation, and I'd love to see someone tackle the "Hard Problem" of actual attrib…

That is impossible. You might be able to do it if you invented a completely different method of image generation, but the amount of original images present in a diffusion model is 0% with reasonable training precautions, and attributing its weights to particularly any of its input is nearly arbitrary.

(Also, it's entirely possible that eg a model could generate images resembling your work without "seeing" any of your work and only reading a museum website describing some of it. Resemblance is in the eye of the beholder.)

Re: Adobe Firefly: AI Art Generator

#343

Post by danShumway »

O__________O wrote: post #214
superbatfish wrote: post #109

Earlier quoted context omitted.

There are at least two potential issues pertaining to copyright law here, and it's not clear which one you're asking about. That's why the responses you're getting here seem to be answering different questions. 1. Are the AI systems violating the copyright protections of the images they were trained on? If so, are users of such AI systems also in violation of those copyrights when they create works derived from those…

As I am sure you’re aware, already posted response US Copyright’s ruling related to authorship here, so will not be repeating myself: https://news.ycombinator.com/item?id=35247377 Will say that post you linked to also states, “17 U.S.C. 101 (definition of “compilation”). In the case of a compilation including AI-generated material, the computer-generated material will not be protected outside of the compilation.” — t…

> a prompt is somehow a recipe and not a novel expression that merits copyright

People keep bringing up photographs, I think the better analogy is commissions. And in fact, the copyright office points towards commissions in its explanation of its policy.

Under current copyright law, if I work with an artist to produce a commission by giving that artist repeated prompts, pointing out areas in the image I'd like changed, etc... I don't have any claim of copyright on the artist's final product unless they sign that copyright over to me. My artist "prompts" are not treated as creative input for the purpose of copyright.

I would love to hear an argument for why prompting stable diffusion should grant copyright over the final image, but prompting a human being doesn't grant copyright over the final image. Directing an artist is just as much work as directing an AI, and in many ways will put you much closer to the creative process and will give you more control over the final product. You can direct an artist in much more specific detail than you can direct stable diffusion. You can be a lot more involved in the creative process with a human artist. And just like with an AI, if you take that artist's final drawing and do your own work on top of it, you can still end up with something that's covered by copyright.

But despite that, we've never assumed you intrinsically get any copyright claim over the artist's final picture that they give you.

So the "prompt as a recipe" analogy seems to hold up pretty well for both AI generators and human "generators". All of the same questions and tests seem to apply to both scenarios, which makes me feel like the copyright office's conclusion is pretty reasonable: prompt-generated art isn't copyrightable, but prompts may be protected in some way, and of course additional modifications can still be covered.

Yes, there's grey area, but no more grey area than already exists in commissioning, and the creative industry has been just fine with those grey areas in commissioning for a long time; they haven't been that big of a deal.

Re: Adobe Firefly: AI Art Generator

#344

Post by novok »

wwweston wrote: post #318
krisoft wrote: post #201

Earlier quoted context omitted.

I'm not a lawyer and I don't work for Adobe. :) The contributor agreement linked from here[1] is this: [2] "You grant us a non-exclusive, worldwide, perpetual, fully-paid, and royalty-free license to use, reproduce, publicly display, publicly perform, distribute, index, translate, and modify the Work for the purposes of operating the Website; presenting, distributing, marketing, promoting, and licensing the Work to u…

I'm starting to think that use of works in a training set is a category not covered well by existing copyright law, and it may be important to require separate explicit opt-in agreement by law (and receipt of some consideration in return) in order to be considered legitimate use. The vast majority of copyrighted works were conceived and negotiated under conditions where ML reproduction capabilities didn't exist and n…

If this kind of abstract copyright regime of 'I had the idea first, and anyone who uses a derivative of my idea must pay me money!' is a very sillpery slope of anyone who makes art or music of any genre needing to pay a royalty to sony/disney, because that is where these 'flavor copyrights' will end up going. The right kind of ambitious amoral lawyer in a common law regime will leverage an AI royalty law into a generic style copyright law because that is what will be needed to write this law properly.

And on top of that, it will become a spotify where each creator gets a sum total of $0.00000000001 per AI their media item was trained on and maybe a few dollars a month, while paying a greater tax to apple-sony-disney whenever their AI style recognizers charge you a royalty bill for using whatever bullshit styles it notices in your media items.

Copyright should stay in it's 'exact duplication' box, lest we release an even worse intellectual property monster on the world.

Re: Adobe Firefly: AI Art Generator

#345

Post by astrange »

jokethrowaway wrote: post #171
mesh wrote: post #124

Earlier quoted context omitted.

>That is, these companies are largely not doing the hard part, which is creating and training these models in the first place. fyi, for Adobe Firefly, we are training or models. From the FAQ: "What was the training data for Firefly? Firefly was trained on Adobe Stock images, openly licensed content and public domain content, where copyright has expired." https://firefly.adobe.com/faq (I work for Adobe)

Sure but are you standing on the shoulders of Stable Diffusion or not? Fine tuning Stable Diffusion with your own images is way easier than creating Stable Diffusion in the first place. If you're creating your own I stand corrected and that's some serious investment.

Stable Diffusion 1.x isn't original work either; it uses OpenAI CLIP.

But training your own is pretty doable if you have the budget and enough image/text pairs. Most people don't have the budget, but at least Midjourney and Google have their own models.

Re: Adobe Firefly: AI Art Generator

#346

Post by astrange »

egypturnash wrote: post #236
teaearlgraycold wrote: post #200

Earlier quoted context omitted.

Is there a license that exists that you could put on your work to prevent its use in model training?

Not as far as I know. There needs to be one, and internet-scrapers need to be able to be sued for ludicrous amounts of money if they violate it, IMHO. Training AI models feels way outside the scope of what I think "fair use" should cover.

Non-commercial internet scraping for model creation is explicitly legal in the EU; the result of a model trained on a billion images really has nothing to do with anyone in particular's art. Although the model would likely work pretty well without ever seeing any "art" images.

Re: Adobe Firefly: AI Art Generator

#347

Post by astrange »

waboremo wrote: post #143
grondo4 wrote: post #135

Earlier quoted context omitted.

> i.e. the act of reading may cause accidental infringement when implementing the "things you learn" Surely you know this isn't the case right? Maybe you're confused because we're talking about programming and not a different creative artform? Great artists read, watch and consume copyrighted works of art all day, if they didn't they wouldn't be great artists. And yet the content they produce is entirely there own, f…

Artists get into trouble all the time for producing works very close to something that already exist. That's like the number one reason artists get shunned in the communities they were in.

The process of online artists shaming each other doesn't really have anything to do with the legal system, though they all act like it is.

Re: Adobe Firefly: AI Art Generator

#348

Post by Thorentis »

NaN1352 wrote: post #253

What’s interesting to me is how this only works because of prior art. BUT, when prior art will be AI-assisted if not 99% generated art, from a pool of prior human art ever so slowly diminishing… where is this going? For one, "art" can only lessen in value. Perhaps physical art will grow in value as digital art’s "made by without AI" tag becomes unprovable and meaningless. I think it’s bad. Whomever provides these too…

As an accelerationist, I can see an upside. Human culture has been in decline for decades, with mainstream art (of all kinds) rapidly declining in creativity and value. There are always exceptions, but I think this was and is the trend.

This AI trend will turn our attention back to what it truly means to be an artist. From the muddy waters of AI art will shine the true works of art that only humans are capable of producing. This will raise the barrier to entry and increase true arts value. This imo will be a good thing.

Re: Adobe Firefly: AI Art Generator

#349

Post by AJ007 »

musicale wrote: post #330
wwweston wrote: post #318

Earlier quoted context omitted.

I'm starting to think that use of works in a training set is a category not covered well by existing copyright law, and it may be important to require separate explicit opt-in agreement by law (and receipt of some consideration in return) in order to be considered legitimate use. The vast majority of copyrighted works were conceived and negotiated under conditions where ML reproduction capabilities didn't exist and n…

Derivative works or remixes usually require a license. Artists could very reasonably argue that AI-generated images are derivative works from their own images - especially if there is notable similarity or portions appear to be copied. They could also point out that their images were used for commercial purposes without permission and without compensation to generate works that compete with their own. For example, ev…

I propose that it is impossible to prove that any content created after 2022 did or did not utilize ML/AI during the process of its creation (art, code, music, audio, text.) Thus, anything produced after 2022 should not be eligible for copyright protection. Everything pre-2022 may retain the existing copyright protection but should be subject to extra taxes on royalties and fees given the exorbitant privilege.

Though this sounds extreme, enforcing the alternative would break any last remnant of human privacy. It would kill the independent operation of computing as we know it and severely cripple AI/ML research when we need it most: human alignment.

It is possible that a catastrophic event occurs and halts the supply chain of advanced semiconductors in the near future, in which case the debate can be postponed indefinitely.

Re: Adobe Firefly: AI Art Generator

#350

Post by jamilton »

roughly wrote: post #243
mesh wrote: post #226

Earlier quoted context omitted.

We are working on a compensation model for Stock contributors for the training content, and will have more details by the time we release. The training is based on the licensing agreement for Adobe contributors for Adobe Stock. (I work for Adobe)

I would be very, very interested to see a compensation system that took into account the outputs of the trained model - as in, weights derived from your work are attributable to X% of the output of this system, and therefore you are due Y% of the revenue generated by it. It sounds like Adobe is taking seriously the question of artist compensation, and I'd love to see someone tackle the "Hard Problem" of actual attrib…

I think the naive approach of just dividing revenue equally across all contributors could be acceptable, and would have lower overhead costs.
Post reply on HN