> fully-generated content is public domain and copyright can not be applied to it.
Some people keep saying this but it seems obviously wrong to me.
At least in the United States, “sweat of the brow” has zero bearing on whether a work is subject to copyright[1]. You can spend years carefully compiling an accurate collection of addresses phone numbers, but anyone else can republish that information, because facts are not a creative work.
But the output of an AI system is clearly not factual! By extension, it doesn’t matter how little work you put in—if the work is creative in nature, it is still subject to copyright.
Sweat of the Brow is irrelevant. Only humans (or collections of humans) can create a work that gets covered by copyright. Non-human animals cannot create copyrighted works, even intelligent ones. Humans can apply sufficient creative transformations to non-copyrighted works to create copyrighted works.
A human needs to have significant creative choices involved in the creation for a work to be copyrightable. Naruto v. Slater litigated this, where a "selfie" taken by a macaque was ruled to not be eligible for copyright.
U.S. Copyright Office, Compendium of U.S. Copyright Office Practices § 306 (3d ed. 2021)¹ is quite explicit:
> The U.S. Copyright Office will register an original work of authorship, provided that the work was created by a human being.
> The copyright law only protects “the fruits of intellectual labor” that “are founded in the creative powers of the mind.” Trade-Mark Cases, 100 U.S. 82, 94 (1879). Because copyright law is limited to “original intellectual conceptions of the author,” the Office will refuse to register a claim if it determines that a human being did not create the work. Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 58 (1884). For representative examples of works that do not satisfy this requirement, see Section 313.2 below.
This has not yet been litigated to conclusion, but it seems likely to me that LLM-generated outputs are not subject to copyright.
> * fully-generated content is public domain and copyright can not be applied to it.
Simpler yet - and inevitable, on sufficiently long time scales - is to dispense entirely with the notion of intellectual property and treat _all_ content this way.
This would remove the incentive to generate content, no? Copyright duration could be much shorter, but I think artists, writers, etc. would prefer the continuing protection of their work. (And I'm pro-copyright reform.)
I'm a full-time professional musician, and I don't know anybody (at least in bluegrass) who thinks that the system of IP is designed to protect us, or is in fact serving us in economic terms. It seems much more geared to protect spotify and apple than it does the musicians.
Last year, I cut Drowsy Maggie with David Grier (something about which I boast every chance I get :-) ), and part of our journey was listening to aging, nearly-forgotten versions to find melodic and harmonic ideas to harvest and revive. For this, we of course made heavy use of archive.org's Great 78 project - and at the very same time, the RIAA (who is supposed to represent us?!) was waging aggressive lawfare against the Great 78 project, to try to take it down.
It was just the height of absurdity.
Consider that since at least 2020, every grammy winner in both the bluegrass and americana categories (and almost no nominee) has been released DRM-free. And that many of the up-and-coming bluegrass and jam bands are now releasing all of their shows, directly off the board, licensed with creative commons-compatible licenses.
The only leverage you have to stop Spotify from taking your music and publishing it without your permission is your copyright of the music.
In fact, every time I see a complaint about copyright it's always "we tried to do something at small scale for some noble purpose and couldn't because of pesky copyright laws," and it completely ignores the massive scale of abuse for profit purpose that would occur if copyright didn't exist.
Think of how AI scraped everyone's books without permission using the flimsy excuse that it's transformative work, except they wouldn't even need that excuse or the transformation. Amazon could just take everyone's books and sell it on Kindle, then kick out all authors because they only need to buy 1 book to resell it as if they were the owner of the book.
> The only leverage you have to stop Spotify from taking your music and publishing it without your permission is your copyright of the music.
There are a lot of challenges facing a band, including the frustrations of CDBaby and Distrokid. If you told me my music would just magically appear on Spotify without my having to lift a finger (and without having to implicitly endorse them by putting it there), that'd be a huge relief.
> it completely ignores the massive scale of abuse for profit purpose that would occur if copyright didn't exist.
"abuse"? If you can somehow make money by playing music that I've made, nothing will make me prouder. And whatever you're doing that's generating that profit, it will almost certainly increase the likelihood that I can plan a series of shows around it, which will in turn generate income for me. Who exactly is losing here? Where is the "abuse"?
> Think of how AI scraped everyone's books without permission using the flimsy excuse that it's transformative work, except they wouldn't even need that excuse or the transformation.
I'm already sold, you don't have to keep making it sound sweeter and sweeter.
I find your arguments hard to relate to, given that I buy CDs/on Bandcamp because I want to pay musicians. I usually can't afford concerts, so I won't be there as a channel for merch. If you release all your work under permissive licenses, how do you expect to be supported?
In my mental model, participating in bandcamp (and getting your supporter badge) is a sort of "merch". You aren't really buying the music - the music is something you hear as the result of a FLAC file being decoded, and that FLAC file can be endlessly and freely copied.
Even better than the bandcamp model, in my thinking, is for you to get the music via pirate channels like bit torrent or IPFS, listen to it, and if you want to, you can buy album-related merch on the ethereum blockchain. That's the long-term vision.
As far as affording concerts: on sufficiently long time scales, I'm wanting to make our shows free-to-enter, buy-a-ticket-stub-if-you-want. Sadly, a lot of great rooms around the country are locked behind contracts with Ticketmaster/AEG and are prohibited from hosting such a show.
> If you release all your work under permissive licenses, how do you expect to be supported?
I believe that a huge majority of our fans are like you: they _want_ to support us, and the spotify model doesn't really give them a channel for that. Permissive licenses don't prevent people from supporting us on bandcamp and similar models.
And it's far from a foregone conclusion that what you call "their right" is real - I don't believe I have a right to stop you from copying bytes on your device. That's insane.
And yes, I'm sure that the top .01% of pop musicians for whom the system is working well will hang on, and many more hoping to hit whatever lottery they've hit.
But as I pointed out above, some genres which are experiencing thunderous revivals right now are embracing DRM-free very hard, and even CC as well.
As John Perry Barlow said about his band, The Grateful Dead, when he found himself facing down record labels and movie studios as the only person on the stage who was actually in a band:
"We gave away our so-called intellectual property and became the most popular band performing band in the United States. and we're making one hell of a lot of money giving it away. It was not required that we be absolutely firm in our hold on this material because we recognize something important which is that in an information economy the normal sense of an economy based on scarcity is turned on its head. Value in an information economy is based on familiarity and attention. These are very different principles and and trying to optimize towards scarcity as you are by all of your methods is not going to be in the benefit of creation."
It's true, before copyright existed, no one made any art at all, and they certainly weren't paid for it. Thanks to copyright, the large majority of artists have been well and fairly compensated for their work.
Okay, you're right. I mean, there was patronage for a long time, and then a good era of proper copyright protections. The modern system really does need a reform, I agree. But I don't think we should wholesale put everything in the public domain. I mean, AI scrapers already think that's the case, but…
It's not just AI scrapers, it's the entire concept of the internet.
The internet _wants_ to copy bytes. That's what it does. Right now you're reading this because bytes were copied from my local machine to the HN server, and then to a cache, and then to your machine.
Copying bytes is, in some low-level sense, the entire function of interconnection in the human species.
The idea of trying to bolt-on little machines to restrict the flow of bytes at every vector of network connectivity, just to satisfy some abstract claim of "property" - it's completely nuts. It's never going to work. Tomorrow it will work worse than it does today, and every day going forward until states realize that the laws they want to enforce around this concept are simply not possible on the internet.
And you know who will celebrate that? Musicians like me. Nobody will be happier to see a lubricated copying machine become the human identity more than people who are trying to get their music out there, and trying to be inspired by the music of others.
Well, what prevents you from doing that right now? The threat that I'll call the cops on you? Is that really how we want the internet to work? It's sure as hell not how I want my music to be perceived - I can't fathom wanting the state to intervene because some kid listened to something I made without permission.
You are welcome to "steal" anything I've ever made if it pleases you. And encourage your friends to steal it from you. If this process keeps repeating, look me up and let's book a show in your area, and we'll play our music _and_ demo our source code _and_ get you all dancin' and trippin' and having a merry old time.
Sorry about the OT, but I see this more and more often:
Project License: MIT
Readme.md: "Project is about 80% Vibecoded.
80% of your project is public domain.
I'm not saying don't use AI. But at least shut up about how much.
This project uses MIT (which IMHO is already a small problem because it has no patent grant), but I see people vibecoding almost all of a project and then using AGPL.
Doesn't work like that. AI code can't be copyrighted. No copyright => public domain.
Please don't sleepwalk only to wake up to some company closing your projects and laughing at you.
Humans applied eugenetics to dogs, animals and plants for thousands of years.
We don't have a single, all-equal dog race. Diversity will remain and might even explode.
At the same time, governments and cultures have tried to tightly control, or heavily "incentivize" who gets kids, and that has led to dark places.
I think we are just entering a phase of "the same, but with tech". Changes are going to be faster, and thus we will notice them a lot more, but IMHO the result is not dependent on tech, but on culture.
> If every dentist did this, the genepool would improve.
Not really, especially since this is widely regarded not as much of a genetic problem, but as a difference in diet of modern society.
I am also unsure if eugentics *necessarily* brings monoculture. We did it for hundred of years to dogs, and while some races are definitely worse off than others, we literally created more than any here care to remember, and many absolutely love races I find truly ugly.
So the problem with eugenetics lies in understanding what culture lies behind it, imho. While there is a pull to uniformity, people don't like too much of the same, because instinctively you understand it loses value. No difference == no worse, but == no better too.
There's definitely a "too big to fail" thing going on here given how many billion/trillion dollar companies around the world now have 18+ months of AI-assisted code in their shipped products.
Several of the big LLM vendors offer a "copyright shield" policy to their paying customers, which effectively means that their legal teams will step in to fight for you if someone makes a copyright claim against you.
* quote: "No court has recognized copyright in material created by non-humans".
The problem now becomes how much AI work is influence and what about modifications
* Courts have recognized that using AI as reference and then doing all the work by yourself is copyrightable
* AI can not be considered "Joint work"
* No amount of prompt engineering counts.
* Notable: in case of a hand-drawn picture modified by AI, copyright was assigned exclusively to the originally human hand-drawn parts.
Notable international section:
* Korea allows copyright only on human modifications.
* Japan in case-by-case
* China allows copyright
* EU has no court case yet, only comments. Most of the world is in various levels of "don't really know"
After 40 pages of "People have different opinions, can't really tell", the conclusion section says "existing legal doctrines are adequate", but explicitly excludes using only prompt engineering as copyrightable
> Am I now assured that the copyright is mine if the code is generated by AI? Worldwide? (or at least North America-EU wide)?
Not only is the answer to that no, you have no guarantee that it isn't someone else's copyright. The EU AI Act states that AI providers have to make sure that the output of AI isn't infringing on the source copyright, but I wouldn't trust any one of them bar Mistral to actually do that.
It is going to be so interesting now that most software is going to be public domain. It's going to be us and the fashion world working just fine without intellectual property rights.
Products directly generated by a generative model are not copyrightable in the US. And therefore public domain. Not a lawyer, but I think the cases and commentary have been pretty clear. If you make significant human contribution / arrangement / modification / etc it can be copyrighted.
Long story short, you can't prevent anyone from using AI slop in any way they want. You would have to keep the slop as a trade secret if you want it to remain intellectual property.
"It concludes that the outputs of generative AI can be protected by copyright only where a human author has determined sufficient expressive elements. This can include situations where a human-authored work is perceptible in an AI output, or a human makes creative arrangements or modifications of the output, but not the mere provision of prompts."
Anyone seen clarity anywhere on what that actually means, especially for things like code assistance?
> Has the situation changed on AI code legally speaking?
lol,
l m a o,
essentially people who use LLMs have been betting that courts will rule on their favour, because shit would hit the fan if it didn't.
The courts however, have consistently ruled against AI-generated content. It's really only a matter of time until either the bubble bursts, or legislation happens that pops the bubble. Some people here might hope otherwise, of course, depending on reliant they are on the hallucinating LSD-ridden mechanical turks.
That's different. Courts have ruled on AI training data as being fair use, but whether you can copyright AI-generated content is another issue.
It's pretty unclear to me where this stands right now. In the USA there are high profile examples of the US copyright office saying purely AI-generated artwork isn't protected by copyright: https://www.theverge.com/news/602096/copyright-office-says-a...
But there's clearly a level of human involvement at which that no longer applies. I'm just not sure if that level has been precisely defined.
> The courts however, have consistently ruled against AI-generated content.
No, they haven't consistently “ruled against AI generated content”.
In fact, very few cases involving AI generated content or generative AI systems have made it past preliminary stages, and the rulings that have been reached, preliminary and otherwise, are a mixed bag. Unless you are talking specifically about copyrightability of pure AI content, which is really a pretty peripheral issue.
> It's really only a matter of time until either the bubble bursts, or legislation happens that pops the bubble.
The bubble bursting, as it is certain to do and probably fairly soon, won’t have any significant impact on the trend of AI use, just as the dotcom bubble bursting didn’t on internet use, it will just represent the investment situation reflecting rather than outpacing the reality.
And if you are focusing on an area where, as you say, courts are consistently ruling against AI content, legislation is unlikely to make that worse (but quite plausibly could make it better) for AI.
Which came from the US sanctioning Huawei because Huawei was making better chips than Qualcomm and Cisco.
US was posturing to ban China access to Arm. Ultimately, it led to a ban on using TSMC for Chinese companies instead. There is no real justification to do a blanket ban on all Chinese companies except a state-sponsored way to to slow China down in AI race.
The Arm China CEO went rogue and spun it off as its own company. ARM HQ were unable to fire him, as he had physical possession of the company's seal stamp. Reading between the lines, the Chinese government chose not to intervene for multiple years.
This stuff constantly happens to foreigners in China and just seems to be mostly due to having bad local legal teams. If your CEO can run off with the company stamps and screw you over.. then it just sounds like amateur hour and you have no idea what you're doing.
The legal system there is fundamentally very bureaucratic - there are rules, but they're very different from the West. You need local help - and a lot of it. You see similar bureaucratic insanity in Japan, though I'm guessing there is just a lot more legal infrastructure for guiding foreign companies there.
If it’s constantly happening, does that sound like an attractive market to work in? I’m very open to the idea of companies screwing up, but day-to-day operations shouldn’t look Kafkaesque.
When you run your business in China, China runs the copy of your business for you ;)
I do understand that we want to try to see "the same" in the stupidity of our politicians that let all of this happen just like that for many years, but we are different.
I can see what both of the above commenters are saying. Here is my synthesis:
In the US, the powers-that-be are often content to let markets, popular forces, and regulation shape foreign companies. In China (I'm no expert, please weigh in [1]) it seems that the CCP is very motivated to make foreign firms serve its industrial agenda while staying under Party control. That usually means insisting on Chinese ownership stakes or joint‑venture structures, so the state always has a foothold in the business.
In this way, politicians of both countries do find ways to "get what they want" from a foreign business -- even if they go about doing it differently.
[1] I'm not ignorant of geopolitics; I do read about China, but focusing on it is not part of my job nor education.
In this situation I disagree. This is a blatant seizure of the control of a foreign owned company for political reasons. This is straight out of China's playbook.
I agree that the West is in general different, as in this is more an exception than the rule. But being in the semiconductor industry, I'm fed up with the stupid rules we are dealing with since the 1st Trump admin. Even more stupid is the US foreign policy affecting EU companies much more than US companies.
The first is not hackable and the second promised much and delivered just an initial dump with dubious sdk license