https://mwhlawgroup.com/can-copyright-in-software-extend-to-...
Forget AI for a minute. Let's say I write a program to generate cool images using Penrose tiles or space filling curves or some other algorithmic process, maybe deterministic, maybe (pseudo-)random. Under that ruling, it sounds like all of the outputs of my program would also fall under the copyright of the original program, unless the creator could prove that they did substantial work in addition to what the program did.
Which then leads to the next question of: what is the "lion's share"? Does it matter if I randomly punch in some parameters and get a nice result on the first try, or if I spend hours trying different parameter combinations to find some thing that looks the way I want it to? Even if the outcome is still considered owned by the software author, is my particular parameter choice copyrightable by me, being my own creative output?
If you think about it, the modern generative "AI" systems are more or less the same thing, but where the input parameter space is something like the entire space of written language, rather than a couple drop-down menus with algorithm settings and a seed for a PRNG. Does it matter if I slap in a simple prompt and get a nice result, or if I spent hours engineering a perfect prompt?
If anything, it should be unambiguous that the prompts are copyrightable just like any other written work, being actual free-form human-readable text. So at least AI artists have that going for them.
It will be really interesting to see what happens as more court rulings unfold.
I'm being facetious of course. I suspect this technology is going to rapidly push us into radically rethinking how copyright works, but it'll be delayed some 20-30 years by incumbents desperately clinging to their hordes of existing IP.
First, there's a bunch of (non-creative) people that just want free creative assets. They don't like it, when creatives insist on being paid for the [considerable] time, effort, education and experience that goes into the creative asset.
Those people are jumping for joy.
Then, we have the creatives, themselves, who may have attended the correct schools, and have the appropriate education, to produce quality assets. They don't want people to just take their work. This looks like a side-jacking. If someone sees a nice picture that was done by a creative, then they can prompt MidJourney to produce one that looks a lot like it, for the cost of the MJ subscription. They don't particularly care (yet)[0] about copyright.
Then, we have the portfolio holders, who are often not creatives, but still hold the copyrights on a lot of creative work. Think Disney, or Universal. Their portfolios are an investment. They don't want to lose that investment.
Then, we have the crooks, who want to make cheap, crappy copies of creative assets, regardless of who owns the copyright. At the moment, there is an element of risk, as they could get caught and punished.
They are also jumping for joy.
[0] I said "yet," because one of the problems with a "wild west" copyright and trademark atmosphere, is that it is extremely easy to make "almost-like" copies, and if you don't protect your IP (or can't), then it is "up for grabs."
This is done now, but it's illegal, and can be stopped; especially if trademark is involved. I was at a Chinese restaurant, once, and was listening to the music. I noticed that there were many strains of hugely popular (and copyrighted) songs, like Scarborough Fair, and Moondance, all mashed together into a continuous stream. The owner confirmed what I suspected. He subscribed to a service that sends legal "muzak" to his restaurant. He had been mugged by ASCAP, once, and never wanted to go through that again.
I've noticed that "free everything" people often insist that everything except their own work should be "freed."
Would you consider a 3D artist generating a texture for his model to be a non-creative? What about a concept artist who mashes together AI generated images? How would you classify Corridor Digital’s Anime Rock Paper And Scissors series?
To me it seems like there’s legions of creative people who are excited to use AI in their works.
I was talking about a coder that can’t draw, using assets for icons and backgrounds, or a CEO that can’t draw or program, using AI for their entire company. I already know of examples of both.
I’m an artist[0]. When I was younger, I even considered making a living at it. I haven’t done any real artwork in decades, though.
But I’m not a particularly skilled graphic designer. I know enough to be dangerous, but most practicing designers can run circles around me.
3D artists are pretty awesome[1]. I think they will benefit from AI.
And doesn’t reward them for their contributions, thus, removing the “feeder” creativity. Soon, schools stop teaching art, because no one wants to become one, and the algorithms end up using the output of other algorithms as their input, and original creation dies.
Original creative output can be ugly and disturbing, but it is original human creation.
Tools can help this (my example is 3D modeling. When the tools were still “raw,” then only engineers could produce 3D art, and it showed. Once the tools and standards progressed, real trained artists started to use them, to produce awesome creations).
I hear, all the time, non creative people drooling over not having to pay or deal with creative folks.
This is nothing new. Just a new venue for typical human greed and laziness.
We’ll figure it out, and, I guarantee, some creative folks will figure out how to use these tools to become quite successful.
And those non creative folks are still gonna have to pay them.
This implies that without a support structure, people are not interested in creativity. Which is absurd. People are still interested in creating even if they're not compensated for it. Many people are proactively creative at great personal expense.
People painted long before it was taught in a formal setting, and people will continue to paint even if they can't use it to pay the bills (as - quite frankly - most painters do)
I know of what I speak. BTDT, got the T-shirt. One of the issues with noncreatives, is they have this illusion that creatives don't work hard, or don't "earn" their money (which can sometimes, be a great deal).
I run into that attitude on a regular. I'm routinely asked to do $20,000 worth of work for free, "because it's fun for me."
I suspect that a lot of us have dealt with that. Development of software and hardware is a creative (and valuable) effort, but people sometimes think we don't work hard at it, because our fingernails aren't cracked.
In my career, I've worked with some of the top people in the design field (and my company paid them a lot of money). They were worth every nickel.
One of the problems with being good at something, is that we make it look easy, so people think it will be easy.
I posted this a while back:
There's the old story I've heard about the retired engineer (apocryphal, I'm sure):
This engineer worked for a corporation for 40 years, and retired.
A few months into his retirement, his old company frantically calls him, begging him to come in, and help fix an issue with the system he worked on. Apparently, the new team had managed to hose it, and couldn't figure out how to fix it.
He comes in, sits down at a terminal, looks at it for five minutes, and says "Here's the issue. If you do this, it should be fixed..." He then presents an invoice for $10,000.
The beancounters go "There's no way we can pay $10,000 for five minutes work! Itemize it!"
He sits down, scribbles a bit, and presents an invoice that says:
Time to fix bug: Five minutes. $20
Six years of college, and forty years of experience, so I can fix a bug in five minutes: $9,980
There's an unfortunate culture of "Everyone [but me] should do work for free, and publish it on the Internet, so I can use it, and make money [because my own work will never be free]."
perhaps that ruling could be justification.
https://www.openculture.com/2020/03/every-melody-has-been-co...
Hell no. I can't tell if you're serious about this. You know prompt is sometimes literally a single English word right? Like literally just "pig"? If that's copyritable then it's officially the end of copyright.
Actually that would be interesting hack for older people. As it could give longer period of protection than their natural lifespan.
This ephemeral fake world of contemporary society is not some sacrosanct construct, or divine mandate. There is no immutable obligation to preserve it.
This case is fundamentally not about whether art in which AI is used as a tool by a human (even if the human role is entirely in programming the AI) is copyrightable as a work of the human, it is solely about whether a piece of software can be an author under copyright law, a very different question.
Traditionally code would be written as a tool to achieve a specific output. If you automate the task of generating a particular piece of code, you still did all the work, or at least defined it pprecisely. While I don't see AI as an independent mind (yet) it seems independent enough from the person who prompts it to say the prompter doesn't own the output. I'm so 100 percent on what I just wrote, but it seems reasonable enough to say the precedent you mention may not apply.
Even if you consider the user input to be marginal, the issue from the link is the copyright of the program extending to the code, not that of the user.
https://www.federalregister.gov/documents/2023/03/16/2023-05...
This is of course up for interpretation by a judge (though I think the US copyright office refusing registration may make it de-facto impossible to get it in front of a judge, as I think it's a requirement to actually bring a suit against someone for infringement), but I suspect this is the kind of distinction that they will consider regarding this particular precedent.
The copyright office refusing registration counts as copyright registration for the purposes of register-before-suing-for-infringement.
I don't see any inconsistency in this ruling.
The code has no meaning without a compiler or interpreter.
The compiler and interpreter have no meaning without a CPU.
If I store x86 CPU instructions in a Postgres database, is that data or code?
It’s also a worth viewing this in light of different computing models. The main mechanism may look quite different if you’re looking at a CPU, GPU, or DSP model of computing.
I agree that both code and data are akin in terms of meaning. Transitioning from Assembly language to Python liberates us from concerns about registers and low-level intricacies. Python directs our attention to the problems we aim to solve, in contrast to low-level languages where a significant effort is spent on specifying hardware operations. High-level languages emphasize outcomes, enabling us to focus on the result rather than meticulously instructing the hardware. AI propels us further along this trajectory.
In this sense, I view a prompt and code as rather similar.
However, code is just data for a compiler.
Data has no meaning without spec. It's a quite big difference. Code has no meaning without spec too.
var a = 1 + 1 << 2;
Using a Gödel like encoding every program is also a number. So it seems that numbers that are complex enough that they are isomorphic to a copyrightable program should be copyrightable. I tend to think of the neural net weights as a different instruction set, just how Intel and ARM are also different instruction sets.
I fully well realize that lawmakers and the judiciary will be the final arbiter here (although I hope some logic will be part of the process).