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

An IP attorney’s reading of the Stable Diffusion class action lawsuit

katedowninglaw.com

Post reply on HN
Print view View original

191–200 of 337 posts

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#191

Post by jeroenhd »

Karunamon wrote: post #38
rafale wrote: post #35

I hope the law will converge to this: As a human, I don't need a license to look and get inspired by art. But I am not allowed to feed that same data to a machine as a training dataset without proper authorization from the owner.

I hope the exact opposite. AI, including AGI if we ever get there, cannot be allowed to be strangled in its crib by artificially limiting the information it can learn from in the name of IP maximalism. IP law already goes way too far, the line should be drawn here.

The technology behind AI and AGI does not depend on copyrighted work. If the models are trained on original work, public domain works, or extremely permissively licensed work (CC0, WTFPL) then there simply is no IP conflict.

The use of including copyrighted materials in the trained model was a choice, not some obvious fact about the nature of AI. All of this could've been avoided if the data set did not include unlicensed work in the first place.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#192

Post by rpdillon »

scotty79 wrote: post #48
TylerE wrote: post #33

Earlier quoted context omitted.

You understand that completely kills OSS as a concept, right?

Why would that be? Huge amount of OSS is released under fully permissive licenses.

Those licenses only carry weight because of copyright.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#193

Post by wvenable »

shagie wrote: post #182
wvenable wrote: post #87

Earlier quoted context omitted.

> Machines aren't people. There's no reason why that is the significant detail. Why does it matter? If you can look at millions of images over your lifetime and faithfully reproduce famous works of art by hand, aren't you just as wrong?

Machines can't create copyrighted works. Setting aside the question of "is the model a derivative work", running the program cannot create a work that is copyrighted. Only humans (and not monkeys) can hold a copyright. And thus, the questions are: "is generating a model based on the data set a derivate work" and the unasked question "is asking the model to generate a work in the style of {artist} a derivative work by…

> Running the program cannot create a work that is copyrighted

If think you're going to need more clarity on what you mean by that. Programs are used to create copyrighted works all the time. And machines can and do create copies of other people's copyrighted works.

> And thus, the questions are: "is generating a model based on the data set a derivate work" and the unasked question "is asking the model to generate a work in the style of {artist} a derivative work by the person asking the model?"

My point is you can take the machine or model out of the question entirely. If you learn stuff and then produce something new with what you learned, is that a derivative work? That's already a complex question but it has nothing to do with how you learned it. It depends entirely on the output and has little to do with the input.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#195

Post by Karunamon »

jeroenhd wrote: post #191
Karunamon wrote: post #38

Earlier quoted context omitted.

I hope the exact opposite. AI, including AGI if we ever get there, cannot be allowed to be strangled in its crib by artificially limiting the information it can learn from in the name of IP maximalism. IP law already goes way too far, the line should be drawn here.

The technology behind AI and AGI does not depend on copyrighted work. If the models are trained on original work, public domain works, or extremely permissively licensed work (CC0, WTFPL) then there simply is no IP conflict. The use of including copyrighted materials in the trained model was a choice, not some obvious fact about the nature of AI. All of this could've been avoided if the data set did not include unlic…

Remember that, at least in this country and I believe in all countries who signed onto the Berne Convention, copyright is the default.

If your AI is limited to only training on the paucity of explicitly permissibly licensed/public domain content (and as I think about it more, this would only apply to things that are permissibly licensed without an attribution requirement, which is something there is no meaningful way to do in a model like the one we are discussing) your AI will not be very useful. With that in mind, I would argue that yes, it absolutely is an obvious fact about its nature.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#196

Post by LeifCarrotson »

haswell wrote: post #109
xeyownt wrote: post #101

I don't understand how using an image as input to a model is a copyright infringement. If the image is freely viewable (say you can browse to it), and you just look at it, are you violating any rights? It seems that violation would only come if you would use the model to produce images that are derivative of that original image, the same way a counterfeiter would make a copy of it. Have the skill to copy is not the s…

> If the image is freely viewable (say you can browse to it), and you just look at it, are you violating any rights? The fundamental issue with this line of argument is that it equates the process of human vision and the consequences of that with that of a computer program ingesting that image and the consequences of that . This anthropomorphization seems like a form of deep fallacy when considering the nature and im…

This anthropomorphization happens all the time in the other direction: Software actors are seen reading license plates, wiretapping connections, checking speed limits and red light compliance, monitoring uploads for copyright infringement, issuing takedowns, and otherwise acting as legal entities. Government actors are constantly allowed to do things because those rights would be afforded to an individual policeman or other human agent.

I agree that this is a dangerous fallacy. Something that legislatures and culture have agreed is fine for a human to do - limited by human scaling, memory, and skill - may not be fine for a computer to do.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#197

Post by cwkoss »

simiones wrote: post #76
jerf wrote: post #50

Earlier quoted context omitted.

Ultimately, this is just something that has to be solved with legislation, not a court case. It's too novel a setup for a court case to deal with under existing frameworks. I think one issue is just that of scale. I personally tend to agree that there's something icky with just slurping up literally everyone's content, then producing a tool that will then proceed to put them out of business en masse . But proving tha…

I think it's not going to be that hard to argue that the company is infringing the copyright of those whose images they are using. Especially once the judge is show how similar the output of SD can be to a particular artist's images with the right prompts (proving that SD has memorized a significant amount of those images).

Some artists images just don't contain much entropy though?

If an AI art engine outputs a frame of solid blue, is it infringing the copyright of Yves Klein's solid blue "IKB 79"?

I think that some artists' styles can be accurately replicated without training on any of their work: because the artists' style is generic enough that it can be exhaustively encoded via the works of others.

This seems like a bad test because generic barely-creative works are much more easily generated by AI engines regardless of the source training data. I wonder if we're going to see IP troll style behavior from artists drawing many obvious things so they'll have standing to sue (and negotiate a 'fuck off' settlement) with AI art engines.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#198

Post by threeseed »

pessimizer wrote: post #137
xeyownt wrote: post #101

I don't understand how using an image as input to a model is a copyright infringement. If the image is freely viewable (say you can browse to it), and you just look at it, are you violating any rights? It seems that violation would only come if you would use the model to produce images that are derivative of that original image, the same way a counterfeiter would make a copy of it. Have the skill to copy is not the s…

> I don't understand how using an image as input to a model is a copyright infringement. Look at the extreme case, then. What if that one image is your only input, and your output is identical to it? What if your output is your input reflected over the x-axis? What if your output just crops the input? What if your output is your input cut into irregular pieces and randomly rearranged? Which outputs violate copyright?…

If the input argument were true then what about apps like Adobe Lightroom.

Would they be able to use your photos for Adobe Stock without permission ?

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#199

Post by troyvit »

williamcotton wrote: post #163
troyvit wrote: post #44

Earlier quoted context omitted.

If you want new art you probably want some form of IP. What's the incentive for an artist if at the first whiff of success their output is overtaken and resold by technocrats with machines?

> What's the incentive for an artist if at the first whiff of success their output is overtaken and resold by technocrats with machines? Because when I have access to these tools I will make better art than the technocrat with access to these tools?

But then are you still an artist or are you now a technocrat?

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#200

Post by shanebellone »

haswell wrote: post #154
l33tman wrote: post #132

Earlier quoted context omitted.

The processes seem pretty alike to me (as a neuroscientist and AI researcher). Things will only move on from here, the next generation of these tools won't use a training set of 5B images and complicated month long training procedures, they will allow the "ingestion" of a style by you showing it a single instance once of a target image and it will immediately know the style (just like a human artist would). I'm not p…

What I meant by the content/context of processes was that one is a biological process that includes all of the context and constraints of evolution, while the other is still ultimately a man-made machine, operating with an entirely different set of constraints, ultimately at the direction of other humans. If we could develop literal eyeballs that could look at these images and translate the information the way humans…

Well said.

If you place a human and a computer in front of a painting. A human seeing the painting is a consequence of biology. A computer seeing the painting is a consequence of design.

There's always a distinction between happenstance and premeditation.

Post reply on HN