AI weights are not open “source” • Hacker News Board

Re: AI weights are not open “source”

#161

Post by version_five » Wed, Jul 05, 2023, 6:51 PM UTC

mike_d wrote: post #154

m4nu3l wrote: post #134

Earlier quoted context omitted.

> you can't make these machine things without literally feeding this copyrighted information into them, therefore they do contain a copy. They don't necessarily do. Think about that. You can take some copyrighted material and transform the information contained in it (for instance a fictional book). You can then write a summary. The summary contains information that was present in the original but it has been transfo…

> The summary contains information that was present in the original but it has been transformed and hence it's not a copy. The summary also contains original thought, something is added to it by a human to make it unique. AI models are primarily deriviative. A better example would be: if I take 1,000 different copyrighted works and put them into a ZIP file, does that resulting file violate copyright?

That example is awful, whatever side of the debate one is on

Re: AI weights are not open “source”

#162

Post by kevin42 » Wed, Jul 05, 2023, 6:52 PM UTC

mike_d wrote: post #154

m4nu3l wrote: post #134

Earlier quoted context omitted.

> you can't make these machine things without literally feeding this copyrighted information into them, therefore they do contain a copy. They don't necessarily do. Think about that. You can take some copyrighted material and transform the information contained in it (for instance a fictional book). You can then write a summary. The summary contains information that was present in the original but it has been transfo…

> The summary contains information that was present in the original but it has been transformed and hence it's not a copy. The summary also contains original thought, something is added to it by a human to make it unique. AI models are primarily deriviative. A better example would be: if I take 1,000 different copyrighted works and put them into a ZIP file, does that resulting file violate copyright?

Let's say you take the harry potter books and create a spreadsheet with each word in it as a column, and the number of times that word appears. Would that violate the copyright? I'd be interested in the rationale if someone thinks it would.

Re: AI weights are not open “source”

#163

Post by pxoe » Wed, Jul 05, 2023, 6:54 PM UTC

m4nu3l wrote: post #134

jrm4 wrote: post #64

Earlier quoted context omitted.

These are not bad arguments, but I don't think they're conclusive. I am a lawyer, and I could absolutely see this going the other way. "You can't make these machine things without literally feeding this copyrighted information into them, therefore they do contain a copy. You can see this by when they reproduce, e.g. the "getty images" deal." *this is not legal advice, dangit commenter person below

> you can't make these machine things without literally feeding this copyrighted information into them, therefore they do contain a copy. They don't necessarily do. Think about that. You can take some copyrighted material and transform the information contained in it (for instance a fictional book). You can then write a summary. The summary contains information that was present in the original but it has been transfo…

what is a 'copy'? byte accurate, or 'something with general resemblance'? would a badly compressed "copy" image of a copyrighted material still be 'a copy' or would it be some other thing? would low quality image compression be enough to skirt around copyright claims? image formats and viewers just 'reproduce' an impression of original data from derive compressed data. it is also just 'information that's been generalized by some degree' - for space saving purposes and so on. so, what if image generators could be thought of as a 'very good multi-image compression algorithm' that can output multiple images as well, to a 'somewhat recognizable degree'.

Re: AI weights are not open “source”

#164

Post by immibis » Wed, Jul 05, 2023, 6:54 PM UTC

mike_d wrote: post #157

immibis wrote: post #152

Earlier quoted context omitted.

This has nothing to do with IP laws, and everything to do with baby formula laws. You don't seriously think that without the ability to sell the recipe, nobody would invent a safe and effective baby formula, right?

I'm not sure you fully grasp all the dimensions of IP law. If you have two brands of baby formula, Death brand that kills babies, and OK brand that is perfectly fine, and you start putting Death brand in fake cans labeled OK brand - that is absolutely an IP enforcement issue. The desire of OK brand to protect their brand, and profits, combined with reasonable IP laws allows them to lead enforcement actions and protec…

That is absolutely not the kind of intellectual property that anyone hates.

The kind of intellectual property we are talking about is the one where Death isn't allowed to make baby formula that doesn't kill babies, because OK patented making baby formula that doesn't kill babies and won't give them a license.

Re: AI weights are not open “source”

#165

Post by jstanley » Wed, Jul 05, 2023, 6:55 PM UTC

slowmovintarget wrote: post #128

jstanley wrote: post #17

Earlier quoted context omitted.

It's a very difficult distinction to make. Would you consider a Python program to be data rather than program just because it is text input to the python interpreter instead of machine code for the CPU?

It is not at all a difficult distinction. Weights are literally numbers computed as output. They are not instructions. The semantics of those numbers even when emplaced (loaded) in an artificial neural net is such that they do not execute. They are not instructions. LLM engines and diffusers perform searches where the weights are used to calculate additional output. Is source code, like Python text, data? Yes. All co…

The difference between code and data only exists in our minds. There is no distinction. Both code and data make the computer do things (and, yes, both code and data only make the computer do things if other conditions are permitting, for example if executed with the right interpreter, or loaded with the right type of viewer). Anything that can be expressed as code can be expressed as data, and vice versa.

Re: AI weights are not open “source”

#166

Post by OkayPhysicist » Wed, Jul 05, 2023, 6:55 PM UTC

Filligree wrote: post #159

8note wrote: post #146

Earlier quoted context omitted.

I consider that one super likely, but then using the model to make competing works with one the artists in their own style is a non-fair use derivative work

Style explicitly isn’t copyrightable. It’ll need to be for some other reason.

Your case wouldn't be about style, it would be about specific elements that you posit were memorized and regurgitated by the model. The fact that you're creating art in the same style/medium as the author is what negates the "sufficiently transformative" fair use defense.

Basically, that world ignores the AI model completely. If your resulting work wouldn't be fair use if you directly were working with something from the training set, it wouldn't be fair use if you fed it through an AI model first.

Re: AI weights are not open “source”

#167

Post by DannyBee » Wed, Jul 05, 2023, 6:56 PM UTC

FrustratedMonky wrote: post #40

Are the weights in our brain copyrightable? Might want to get ahead of the curve on this one. How would this work? Would I get a tattoo with a license spelling out covering the contents of my body?

Has to be fixated (unchanging) and in a tangible medium.

Re: AI weights are not open “source”

#168

Post by mirekrusin » Wed, Jul 05, 2023, 6:57 PM UTC

Animats wrote: post #119

ndriscoll wrote: post #33

The complexity described seems to be resting on the unestablished idea that weights are copyrightable in the first place. If they're not, then presumably "available weights", "ethical weights", and "open weights" are all the same: open weights. Either your weights are under NDA and presumably considered to be a trade secret, or they are public, and the words in your "license" mean absolutely nothing? That seems like…

> The complexity described seems to be resting on the unestablished idea that weights are copyrightable in the first place. Yes. Weights probably aren't copyrightable in the US. See Feist vs. Rural Telephone, in which the Supreme Court ruled that telephone directories are not copyrightable. The copyright clause in the Constitution ("To promote the Progress of Science and useful Arts, by securing for limited Times to…

Assuming that weights are not copyrightable, how much restrictions can you put on output through API from those networks/weights?

Ie. if ClosedAI says you can't use output of their API to train competitive models - is that enforceable or not?

Re: AI weights are not open “source”

#169

Post by kevin42 » Wed, Jul 05, 2023, 6:58 PM UTC

stale2002 wrote: post #135

pfdietz wrote: post #105

Earlier quoted context omitted.

Copyright is for things that are the result of human creativity. If the weights come from running an algorithm on a training set (that one does not have a copyright to) then how can the weights then be copyrightable? They might be a derivative work, but that just means they infringe copyright, not that they are copyrightable themselves.

The answer would be if the weights are transformative enough, and the copyright would come from the person who decided what images to include in the training set. The act of choosing to place images in a certain arrangement, such as a collage, can be copyrightable. The same could be said for the "act" of choosing what images to include in a training set and which parameters to use to train the model.

Does that mean if someone copies a phone book but leaves out some numbers and adds some other numbers then it's a creative work?

Re: AI weights are not open “source”

#170

Post by photonerd » Wed, Jul 05, 2023, 6:59 PM UTC

mirekrusin wrote: post #168

Animats wrote: post #119

Earlier quoted context omitted.

> The complexity described seems to be resting on the unestablished idea that weights are copyrightable in the first place. Yes. Weights probably aren't copyrightable in the US. See Feist vs. Rural Telephone, in which the Supreme Court ruled that telephone directories are not copyrightable. The copyright clause in the Constitution ("To promote the Progress of Science and useful Arts, by securing for limited Times to…

Assuming that weights are not copyrightable, how much restrictions can you put on output through API from those networks/weights? Ie. if ClosedAI says you can't use output of their API to train competitive models - is that enforceable or not?

That would be down to contract/terms of service. You’d be in breach of that, not copyright