What does copyright say about generative models?
oreilly.com
oreilly.com
It's rather a rent-seeking industry; the vast majority of artists benefit only marginally from copyright; an original intent, to give composers an income who would otherwise be out of the monetary loop, is long forgotten; instead, early on it was all about protecting the publishers' business by restricting copies; ironically, composers (or musicians in general) today earn best, on average, when they work as a clerk for a copyright collecting society; I don't think patent and copyright law can be fixed; they can only make it even more complicated and unwieldy, so that it gets even further away from composers and authors, and instead plays into the hands of trolls or monopolists.
> fixing copyright law to accommodate works used to train AI systems, and developing AI systems that respect the rights of the people who made the works on which their models were trained
And then also charge each student of art for studying original works with the intent to create new works based on what they have learned to make a living? This idea can be extended in any direction and quickly leads to a system that is rather against an open society where people benefit from each other.
Consider it a black box where copyrighted work goes in and similar work pops out. Conflating what is essentially a function approximator with human beings when it comes to law is a stretch.
If a generated work is similar to an already existing copyrighted work, it is not relevant whether the work was generated by a DNN; the author of the original work can file a claim anyway. On the other hand a work generated by a DNN by itself is not copyrightable, since a human author is required by copyright law. The use of training data for the DNN is fair use according to jurisdiction in comparable cases.
You could get this score alongside the results if you wanted to generate images but know which ones probably shouldn't be used commercially, or you could choose to filter your results by some similarity threshold
That's more of an open question given that the lawsuits concerning exactly this haven't made it before a judge yet nor have concluded.
There are some very sound arguments against this particular case being considered fair use, and they haven't seen the inside of a courtroom to be tested.
I don't really understand the logic here. I have heard there is also some legal precidence to this claim. But surely using photoshop to create an image doesn't mean that you didn't create the image, but photoshop did? So why should this apply to tools that use neural networks?
Are you sure about that? https://www.smbc-comics.com/comic/themes ;)
There are indeed some people who were able to make a good living from the money they received from the copyright collecting society, though they are a small minority; all musicians I know who are members of a copyright collecting society get a few hundered Swiss francs a year and don't live from their music.
Some years ago I analyzed the yearly business report of SUISA, the Swiss copyright collecting society responsible for musicians, and found that of the 120'000'000 CHF payed by the society only one sixth was given to Swiss composers ("Urheber"), and that the probability of any of those composers would get enough money from SUISA to make a decent living is less than 0.1%; for comparison, the probability of being seriously injured or killed in a traffic accident in Switzerland is > 0.2%; the vast majority (> 97%) gets nothing to less than 1900 CHF.
In comparison the administrative cost to run SUISA in the same year was 26 mio CHF (compared to the 20 mio CHF payed to musicians), and the publishers received 54 mio CHF. SUISA payed ~19 mio CHF to their 166 employees; in 2006 357'000 CHF was payed to the director alone.
I'm not sure if it's copyright that has really allowed for music to flourish, I think that's a claim that requires a bit more research behind it.
You sell your music for $1. I sell your music (but without copyright, it's my music, really) for $0.50 in all the same locations as you do. I'll win any race to the bottom, because my cost in time is absolutely less than yours in the effort to sell my music.
It sold >100,000 units[3], and AFAIK no one's massively undercut the author. I guess he's kind of undercutting himself by making it available on github, but I bet most people would rather pay 3€ and have nice steam features than download and compile a game off a site they likely don't know.
And yes, I do realize he's still the copyright holder for SNKRX, but he's not really enforcing it with a permissive license like MIT.
[1]: https://store.steampowered.com/app/915310/SNKRX/
Well, if you don't have copyright, anybody could also publish the game on steam or in any other shop with the same "nice steam features", and get the money over him. It would just be up to the audience charity ("let's give our money to the original author instead of buying the cheaper same featured alternative release")
Nothing, that's my whole point.
>And if nothing is stopping it, why hasn't someone done it yet (that I know of).
Not everything that can happen has happened (or will happen).
But many programs in similar open status were indeed taken and sold this way. And sure, it's legal and a particular creator might be fine if that happened to his program. But that wasn't what we were discussing in this subthread, but the fitability of this model for people wanting to make a living, e.g. the proposition that: "I guess he's kind of undercutting himself by making it available on github, but I bet most people would rather pay 3€ and have nice steam features than download and compile a game off a site they likely don't know".
Well, first, the undercutters might put their version in a nice site like Steam as well. And that's just one case (or e.g. Steam discourages them, even though what they'd be doing is legal), there's also games/apps that they are sold in the author's website or on iOS/Android app stores - and for those anybody undercutting them with their own build/copy would have an equally convenient/prestigious delivery trivially.
Now, there are some cases where the original creator would fare better, but not because of the basic mechanics of the model, but because of other factors: e.g. a creator could have a nice following and be well loved in the community, to the point that most users would go out of their way to buy their version for support, and not the third party ones, even if the latter are undercutting it.
For the average program and the average creator without fans though, if they are concerned about their own sales, that's more of a "put it out in the open and pray you don't get undercut" proposition...
> including without limitation the rights to [...] sell copies of the Software.
The OC can spend their time on the updates, and I'll spend my time on marketing the new version. I'd probably want to report the original as a clone, because I can and it could help boost my own version's rankings. If the OC wants to fight that, it's his time, so cool.
And should the OC actually succeed in getting that clone off the store, the other 10 I've put up while they're busy fighting against the first clone can take its place easily enough.
After all, without IP - without copyright - it's mine to sell however I want to. Nobody can stop me, nobody can even really chastise me because nothing was taken and nothing was lost.
That copyright you're dismissing is worth a lot. It states that it is, indeed, the owners' work. It raises potential legal ownership issues with the compiled work (which isn't under the same license as the code) and store listings. Even if he's not actively using it, it's still busy protecting his work.
Some potential arguments you might use are that it does happen, and will happen eventually to this project, that it's simply a matter of time. Another is that it currently isn't worth it for the amount of work you might put into doing this for this project, but as soon as it is worth something, it will be. Which implies in this case, the cost of all of this activity would not be covered by the return, even if you didn't have to pay the creator.
There's almost certainly a market efficiency in scaling up this packaging activity -- so that the original creator could give up some of his copyrights to pay someone to do the packaging, as long as that packager is doing it for lots and lots of other projects. That's how IP really works.
But in a world of explosive creative access -- where making copies, and creating works is extremely popular and hypercompetitive, the chances are that copyright as a sellable, alienable right isn't going to be the thing that pulls in the money for the artists. It's a thing that maybe you speculatively sell to that large-scale packager, that promoter, in a buyer's market, for a vanishingly small amount. Which makes the contrast between a world without copyright, and a world with, far less diabolically opposed. IP gets you a pittance, but for the vast majority of artists, it's not nearly enough. It may actually be kind of cruel to incentivize artists to produce with an endlessly drawn out false promise of IP riches.
Could we do better? Well, one thing is that we could stop what you're describing without an alienable right to copy, more like a system to prevent the kind of fraud or misrepresentation you describe. There are components of this in our existing creative incentive laws, but they're actually some of the least "property"-like attributes, and do not revolve around making copies -- moral authorship, some elements of trademark. But it involves a lot of work to try and reform our outlook to come out with policies that genuinely rewards artists and enables innovation. I don't see it arising from those praising current intellectual property models, and arguing for their expansion.
Because you cherry-picked an example. This (separate publishing profiting from the same work) has been happening time and again to musicians and authors - even to those under copyright protection -- the only difference is that those have some legal recourse, and have been shutting down those alternate editions.
For example, Miracle of Sound doesn't do live performances, because his music is all electronic and created by one person.
Plus, I can take my music and do "live" performances of it too. I can also hire folks (say, DJ's) to do these "live" performances on my behalf.
Because of it, or in spite of it?
I'll grab some popcorn. This could be one of the all time epic battles.
Earlier someone posted, and then deleted, an Ask HN: "Are artists fighting AI Art repeating Metallica versus Napster?"
This is actually an entertaining question, because it brings in the power of the entertainments business.
Where is Napster today? Didn't Metallica win that one? It might not be a great comparison, but what if RIAA, MPAA, Sony and the game industry decide that "generative AI" occupies the same threat space as "piracy"?
It was of course Metallica and an army of ten thousand lawyers and goonies from a vast, wealthy, moribund industry that actually did manage to block the road of progress and frighten the genie back into the bottle.
In fact, the power of the film and music industry to shape technology has been so immense, you have to wonder whether they could do it again over "AI".
Right now I think the entertainments industry is shitting itself over LLM technology, but is split over whether it can gain enough control to allow it on it's own terms, or mobilise to fight it. We haven't yet reached the stage of commodity proliferation. That will be the watershed.
The end result was a compromise that still funneled some degree of money into the labels, with the artists getting an equally raw deal as before proportionally speaking--but now with a micro share of your $10/mo to Spotify or wherever instead of their share of $10+ for a single album.
I'm not saying the prior business model was sustainable (at least ethically) but at the end of the day, "if you can't beat them, join them" is still one hell of a compromise to make.
No, it will be boring, the outcome is clear: those with the deepest pockets will win this battle.
Just look at Disney who had copyright law changed back in the previous century.
But you merely beg the question. Which pocket is the deeper? The left pocket of a network of industries that stand to dominate the world via AI? Or the right pocket of the same interests who will be destroyed by it?
In such a world, human creators would be paid a pittance by large rightsholders to create, and then those rightsholders would license those creations to be trained on, and then the human would be cut out of the loop for subsequent creations. In this world, creators wouldn't be much different than the data labelers that currently tag ML datasets. They'd probably be paid the same, as well.
* States take a laissez faire approach to regulation these networks, we get an open source community of people developing them publicly for anyone to run on their local machines, run small clusters etc.
* States take a heavy handed approach and prevent anyone distributing a model trained on copyrighted works without agreement from the copyright holder, or paying some form of tax that gets redistributed to them. We end up with an oligopoly of ad-supported, data-collecting, heavily censored SAAS models which we cannot download, modify, or run locally.
There is no world imaginable where the concept of owning ideas should be protected. Imagine if chefs had to license recipes…
The arts communities need to take cues from the scientific community where citations are the cultural norm. And as a society we need to figure out how to protect artists by protecting and celebrating the expression of ideas. Not the ideas themselves.
Copyright does not apply to ideas, it applies to specific expressions/implementations of those ideas, which is what it seeks to incentivize. The fact that many companies (particularly Disney) sued people who only copied the ideas and not the expressions shows a failure of the legal system, not copyright.
> On March 10, 2015, a jury found Thicke and Williams, but not T.I., liable for copyright infringement. The unanimous jury awarded Gaye's family US$7.4 million in damages for copyright infringement and credited Marvin Gaye as a songwriter for "Blurred Lines". In July 2015, the judge rejected a new trial and the verdict was lowered from US$7.4 million to US$5.3 million
The developer of a notable mobile game had to pull some music from their title a couple years ago because the composer they hired blatantly plagiarized some other works, for example.
No, I am not.
I do not believe it is possible to accidentally plagiarize something (it cannot be by definition). I do not wish to live in a society where you can face legal consequences for picking up a musical instrument and simply expressing what's on your mind. Or as another commenter put it, dropping bricks on a piano. It's insane but it's rather familiar, sadly.
It's in fact you who are falling victim to the common misconception that copyrighting the performance of a riff by recording it grants you rights to attack others' performance of the same riff. And that if I perform the same riff you performed I have somehow plagiarized it. To use your words, if you copyrighted the expression of one idea, then how can you possibly attack someone else's expression, that is not your expression, of the same idea, hmm?
I am lamenting the fact that it's even possible to become so confused with the current state of copyright law (and juries do). And that it's even possible to bring a case against someone in the first place (it is). That is a problem and that's what I'm attacking.
As a thought experiment, consider: https://libraryofbabel.info. Should the author be entitle to copyright over every combination of characters imaginable simply because they wrote them down?
This is not a "common misconception". It's how the system works! I'm not saying it should work that way, but that's how it works.
A series of notes is not an idea, it's a composition. That's just how it is. You can have the idea for a composition, perhaps even for a note progression you want to work into it in your head, but once you've made a composition, it is no longer just an "idea". Obviously there has to be a line somewhere, you shouldn't be able to copyright a sequence of 3 notes, but that's a question of law. Are you a lawyer? If so, I'll defer to you on this.
> Should the author be entitle to copyright over every combination of characters imaginable simply because they wrote them down?
Writing down a series of characters is kind of the definition of authoring a work, yeah. So I would expect they have the copyright if they wrote it down, unless someone else wrote them down first. Should it work that way? Maybe not.
You're missing the nuance. You should watch Adam Neely's latest video on music and copyright.
To elaborate slightly, there's the idea of the series of notes, and there's the actual composition of the series of notes, and there's a performance by some artist of the series of notes.
1. The idea of the series of notes is not copyrightable. If it was somehow an invention it might actually be patentable, but intellectual property is not exactly what we're discussing.
2. The composition of the series of notes, written down or fixed to some media in some form, is copyrightable. And the author of the composition owns the copyright as soon as they generate the work.
3. The performance of the composition, if recorded, is copyright of the person making the recording.
Now here's my commentary.
On 1: some people think that the idea of the series of notes is what you copyright when you write the notes down. This is false. But society is incredibly confused on this topic, as the existence of multiple copyright cases regarding this topic demonstrates.
This may be the fault of outdated legal doctrine, influence from large publishing conglomerate, etc., and in reality probably all of the above. But that's besides the point.
On 2: you compose some music and print it out. This is your work. I am not entitled to take that sheet of paper and run it through a copy machine. I must obtain your permission. Even if I buy that sheet of paper from you, I am not entitled to make copies of it unless I have sufficiently licensed the work form you for that purpose.
I am allowed, however, to create an identical composition. I am even allowed to create an identical composition after having seen yours or heard a performance of it. The law doesn't say I can't plagiarize. Society generally does not value plagiarized works, though, because they lack any sort of creativity.
I am further allowed to sell the piece of paper you sold me. You don't own the item for which you hold the copyright after the first sale. This is how used book stores work and it's getting fucked up as we make things digital. I am also allowed to rent the piece of paper out, sharing your work with someone else. This is how libraries work. I own the paper.
On 3: the record of a performance is the property of the entity that recorded it. I'm sure there are plenty of clauses regarding who is allowed to record performances for popular artists and you probably waive certain rights when you purchase tickets and enter into an agreement with the venue, but that's not exactly copyright law either.
I am allowed to listen to F.U.N.'s Carry On in concert and then belt it out in the car on the way home, even if there are passengers and my windows are rolled down and I'm driving through a crowded public street. I can cover the performance of a copyrighted work. I am not generally considered to be plagiarizing because I am creating a derivative work by performing my own rendition of the song.
Now there are formal agencies that try to govern the performance rights of a composition and try to rent seek on all aspects of the copyrighted work and govern how the work may or may not be used in derivatives. And they are successful in doing so. I don't have the legal training to know how much of this is legal vs how much of it is simply "good" manners, but I do know that, to keep the machine oiled, artists generally acquire mechanical use licenses from other artists before covering one of "their" songs.
That's the extent of how the "system" is specified to work today, as I understand it.
> Obviously there has to be a line somewhere, you shouldn't be able to copyright a sequence of 3 notes, but that's a question of law.
The law deliberately does not specify where this line is. Probably because it's impossible to do so.
So the things I'm fussing over:
a) interested parties construing the idea of copyright to include the, well, idea of the copyrighted work, not just the actual manifestation. This is poisonous and the fact that people are confused is an indication to me that the current system is not clear and specific enough and needs work. As a society we cannot thrive when people have been convinced that a sequence of 8 notes is inexpressible and cannot be repeated simply because someone else did it, or simply thought of it, first. This goes against everything beautiful about creating art in the first place and it is beyond silly that we are wasting legal resources entertaining the folly of large record labels and publishing organizations.
b) it stems from (a), the idea that a copyright holder is entitled to full creative control over the use of a fairly acquired copyrighted work. Even if we conceed the unadulterated original spirit of copyright is perfectly fine and the "system works", it does not extend any sort of authority to the copyright holder to govern how their work should be derived, expressed, interpreted, etc. It is insane that the industry norm right now is such that artists feel they need to obtain a special license of some copyrighted work in order to create a derivative work that incorporates it, or even just to perform their own rendition of it. I don't personally find any reason to fundamentally or axiomatically bestow this sort of creative authority upon people who simply come up with ideas and write them down.
c) finally, an attack on copyright itself. It's a dated concept from an age where copying was hard and doing it was a meaningful function in society. We have generated all text strings and melodies that ever were and can ever possibly be. Our tools have advanced to a point where someone can actually write everything imaginable and claim copyright. Our culture has advanced beyond sheets of paper and the media we use regularly makes copies of data all the time in order to even operate. We have developed algorithms that remix content and generate new content that is at times indistinguishable from the original. Is it a copy? That's the question of the article.
I suspect we all agree that humans who pursue creative endeavors are valuable to society and should be supported. I am simply saying that copyright is a very poor implementation of how our current society might do that, not least because it doesn't actually end up supporting the content creators very well in the first place and has since been coopted by self-serving publishing companies for profit. But also because it doesn't fit with our current understanding of a digital society nor does it mesh with the technology we've developed (the web, computers, and content generating algorithms). I don't really wish to see society stifle the possibilities we unlock with "AI" generated content because of dated notions that we need to protect original creators ability to dictate the copying of their work.
Github copilot would (until they manually changed it) spit out the famous fast inverse square function word for word - comments and everything which is more than just the idea (why would you want the comments with them swearing?)
On one side, you have music and musicians, where the written form is somewhat of meaningless tool used to produce creative expression, which is the performance of sound on an instrument.
On the other hand you have authors (and recently, software engineers), where the creative expression as most understand it, is in the choice of arrangement of words (or statements).
You can't really resolve the two. And so I lean towards a stance that it is fraught to try and build a flourishing creative society on the idea that simply writing an idea down (or saving a file of data) means you own it and can legally bring a case against someone else who happens to write the same idea down, be it music or a book.
And then there's https://libraryofbabel.info, which contains all possible combinations of characters that every has been, or ever will be, written. Should that be copyrightable?
Anyway, I am of the opinion that Copilot does infringe on copyright as we know it. But I'm also of the opinion that simply giving individuals universal claim to any written text they produce is problematic. And to answer your question, maybe I do want the comments. Maybe they communicate something that the code cannot, swearing be damned.
I'd prefer a culture of citations for written/stored/saved works. And a culture of celebrating performance of the arts, not storage of them.
I think the is the crux of the issue. I think it would not. IMO the spirit of copyright law protect certain essential elements of a work being reproduced identically or nearly-identically (1), but not a composition that is reminiscent of a general style of some other work, and certainly not some collective body of work. To me the source of the training data should not matter insofar that it does not matter if an artist studies a body of work to create pieces in the style of X. Does it concern copyright law that an artificial rather than biological neural network has been trained on a body of work? Would a copyright judgment ever be made on the basis that an artist studied too much of another person's work, and not simply the similarities between the piece they created and another specific work?
It is against copyright law to release either a computer generated lowfi cover or a human generated violin cover of stairway to heaven. On the other hand you could go and buy an amp, guitar, pedals, etc. to perfectly recreate a classic 'jimmy page sound', and whatever you compose would be fine as long as the song does not borrow specific melodies found in led zeplin songs. See any Greta Van Fleet song. Moreover, that GVF reminds everyone of led zeplin is no coincidence. From wikipedia: "Greta Van Fleet is often compared to Led Zeppelin. Jake said he went through a year of really intensely studying what Jimmy Page did to the point I knew how he thought."
At the end of the day I think there is no basis to claim a copyright over a general "style", and that's a good thing.
(1) https://www.briffa.com/blog/classic-copyright-cases-ice-ice-...
Well, pastiches are a thing, and I've read and watched a few. Never heard about anyone getting into trouble for them.
Is it, though?
The article wrestles with the notion of the gap between "idea" and "expression." To me, I wonder if this is the same gap. The training data is equivalent to the "idea," and the output of using that training data in a particular way is the "expression."
In this view, the result of your training isn't transformative, and it might not even something you can claim copyright over. What is it other than a particular arrangement of facts that have been feed into it? Merely adding weights in a highly dimensional space does not seem "transformative."
This article feels like it's wrestling with the wrong side of the problem.
I don't buy this, given that NNs can and do encode data from their training sets in the network itself, similar to compression, and they can regurgitate that code when given the right incantations. They can also encode "the heart of the work" when it comes to algorithms from copyrighted work being recycled in the NNs output.
And I don't buy that it's transformative, either, given that it's taking someone's code as input and just outputting more code. The purpose is literally the same, building software for profit. It's not like the code is being used in an art project, it's being used for the same purposes that the copyrighted works were created. This is not like making a collage out of magazine cut outs, it's like taking "People" magazine's content and rearranging it and selling it as "Persons" magazine for profit.
When it comes to fair use, transformative derivative work is not the only factor considered. Another factor is the purpose and character of use, especially whether or not the derivative work is used for commercial purposes. Code regurgitated by these models are used for the same thing as the code they were trained on: building software commercially.
Another factor is "the effect of the use upon the potential market for or value of the copyrighted work". It's beyond obvious that laundering copyrighted code for commercial use affects the potential market for and value of the copyrighted work.
But I don't think the artists I'm seeing on Twitter understand that, even in their best case, where training on a copyright image is considered total infringement, this genie still isn't going back in the bottle.
Models are getting better and training more efficient all the time.
Soon you'll be able to train a model entirely on public domain content and have great generative art output.
It might be harder to express the style you want than just sharing a name of an artist in that style, but that won't hold things back.
With this worldview the discussion about whether training on something is infringing its copyright fast becomes irrelevant, for better or worse.
We need IP law reform and UBI/strong social safety net, the problem isn't automation, it's that we allow people to own the automation and charge rent on it.
Singapore version: https://www.twobirds.com/en/insights/2021/singapore/coming-u...
Other countries have already made copyright exceptions for AI training.
This is the SHA256 of Captain America - The Winter Soldier. Pulled from my Blu-Ray copy of it. It can only be produced (in the current universe) if you have access to the complete data from the Australia Blu-Ray release of that movie.
So to create that number, every single frame and pixel of the original movie was included in some way.
Is the number itself copyright? After all, every byte of its input definitely was - there's no way to produce it without those bytes.
For a work to be copyrightable (including derivative works), there needs to be a quantum of creativity. Since there's no creativity in the process of constructing a cryptographic hash, there's nothing copyrightable there.
That's a horrible rule of thumb. The canonical example of a derivative work is adaptation--the film of the book, for example--which frequently increases sales of the original work.
Drawing sales away from the original work is one of the factors used in fair use, and frequently ends up being the most important one.
Copyright doesn't protect skill, as it doesn't protect algorithms, because these are tools to create and not finished products.
Like how Vanilla Ice "stole" Queen's bass line.
As someone who grew up in the 80s and 90s, I also really wish we'd gotten the explosion of almost completely sample-based music we were starting to see with bands like the Beastie Boys, the KLF, and Pop Will Eat Itself. The Biz Markie sampling case basically shitcanned an entire nascent subgenre.
https://www.youtube.com/watch?v=D-uV8TGjaGU
Which sampled the guitar melody from Lou Reed's "Walk On The Wild Side."
https://www.youtube.com/watch?v=oG6fayQBm9w
And the "Rebel Fanfare" from the Star Wars Theme composed by John Williams.
https://youtu.be/_D0ZQPqeJkk?t=278
Which sampled this part of "The Sorcerer's Apprentice" by Paul Dukas. (Okay, not quite but it's easy to guess the inspiration.)
Without samples and sequencing being the primary production technique, suddenly real instrumentation and organic musicality became the thing. Hip hop started building on covered riffs and musical passages more than straight up samples (I’m guessing the licensing was easier).
That also gave us Check Your Head from Beasties, as the follow up to the pioneer sample-laden album Paul’s Boutique, where they famously played their own instruments again and became even more popular. The Native Tongues scene also gave us Prince Paul/Dr. Octagon, which gave us Deltron 3030 and Dan the Automator, which gave us Gorillaz, which…
There really was a lot I suspect would’ve gone quite differently without that court case.
An infringement being accidental also doesn't seem to stop copyright holders from successfully suing people. I think this means you could technically be successfully sued for dropping some bricks on a piano, since no matter where they land, the action constitutes a public performance of a copyrighted work. Fun stuff.
PS: I also second the Adam Neely video. I was actually wondering if someone would link it (:
The disappointing bit: I say it's an abridged version because I distinctly remember him talking about how he actually claimed copyright for all 8-note melodies under a permissive license, which I can't find in the one that actually loads. He "brute forced" every 8-note melody with a program, saved them to a disk and claimed copyright under (I believe) the MIT license. (I can see legal issues with doing that, of course, so it's not hard to imagine possibly why the video was replaced with a different one.)
I may not agree that it should be 8, but I don't see any rigorous or well-reasoned model here to explain why "you can copyright 8 notes" is a festering disease when it seems like people are not reasoning through why something should or shouldn't be copyrightable.
Copyright isn't a law of nature, but a tool to balance needs of creators and everyone else. With this tech the power has definitely shifted, so the existing law may not be adequate any more.
The reason why fair use is vague is specifically to confuse people who ask these kinds of questions. The Supreme Court needed a tool that artists could use to legally smack down people who republish fragments of other people's work, but didn't want to abolish the 1st Amendment in the process. So basically judges have the final say as to whether or not something is novel creativity or in debt to the original. Any hard-and-fast rule beyond "binding precedent applies" is effectively copyright abolition by degrees.
>We lost most of Elizabethan theater because there was no copyright. [..] Without some kind of protection, authors had no interest in publishing at all, let alone publishing accurate texts.
This is a dated example, if only because creative works leave a lot more evidence now than they used to. People today will act to preserve art against the artists own wishes and at great personal risk.
>and it’s easy to suspect that the actual payments will be similar to the royalties musicians get from streaming services: microcents per use
Given the amount of data these systems need (read: more than humanity can provide) I'd say microcents is arguably too high. Remember that you can't actually derive a clear chain of value between one particular training set entry and one particular execution of the model. It's all chucked into a blender that runs on almost-linear algebra and calculus. At best you can detect if parts of the image resemble specific training set examples[0] and pay people slightly more if the model regurgitates training set data.
Let's also keep in mind that a good chunk of the licensing system is based on being able to say no to specific users, or write very tailor-made licensing agreements for specific works or conditions. That's still going to be threatened, even if we can pay sub-Spotify-tier royalties every time a model trains itself on your work.
>It is easy to imagine an AI system that has been trained on the (many) Open Source and Creative Commons licenses.
Working on it: https://github.com/kmeisthax/PD-Diffusion
The thing is, we already have a good database of reusable, public-domain, no-attribution-necessary images; it's called Wikimedia Commons. I really can't fathom why OpenAI didn't start there, other than just an assumption that they were entitled to larger datasets or a feeling that they could get established before anyone sued.
Even then, OpenAI already tried this with computer code and they're getting sued for it anyway, because they never bothered with attribution in the case of training set regurgitation.
[0] This is possible because part of the prompt guidance process involves a thing called CLIP which can do both image and text classification in the same coordinate system.