Not necessarily. The referenced guidance [0] says: "...copyright will only protect the human-authored aspects of the work, which are 'independent of' and do 'not affect' the copyright status of the AI-generated material itself." If you read the paragraph or two above that one, it really seems like products of agentic coding cannot be copyrighted, as there wouldn't be significant authorship involved.
Very glad someone actually read the decision and understood it, despite how much reporting on this has been poor. This was not a case about "can AI-generated art be copyrighted?", despite all the reporting misleading people. (Including me, until somebody finally pointed me at the actual decision — https://www.lexology.com/library/detail.aspx?g=24e0581c-2c28... — and I could read it for myself). The judge literally quoted that case where a monkey picked up a photographer's camera and triggered it, saying that only humans can hold copyright: not animals, and not tools. And he also specifically said that he was not addressing "how much input is necessary to qualify the user of an AI system as an ‘author’ of a generated work".
So it's not the case, contrary to what many people (including me!) have said before, that the decision was "works produced by AI tools cannot be copyrighted". Rather, it's "you cannot assert that the AI tool itself is the author, you must assert that a human is the author". And the amount of work put into the prompt will definitely matter.
In other words, if you just prompt “draw a picture of a cat” then it’s possible you didn’t put enough work into the image to count as the author. But if you have a specific picture in mind that you want to create, and you prompt “draw a picture of a two-year-old cat with orange fur and orange eyes, in a sitting position, looking out of the window of a train. The interior of the train is lit with dim orange lighting. Outside, it is night and there is a full moon visible through the train window,” and then you refine that prompt until the AI produces an image close enough to what you had in your mind’s eye, then that image is clearly your own creation: the AI tool was just the tool you used to take the idea in your head and turn it into an image that other people could look at. Whether you use a paintbrush, a digital-art creation tool like Krita, or a digital-art creation tool like Midjourney, as long as you came up with the concept and did the necessary work to make the tool produce the image, then you're the author and you can assert copyright. (Note that this paragraph is my own opinion, not the judge's ruling, but I think it's a pretty defensible opinion: "draw a picture of a cat" might not be specific enough to assert that you created the resulting image, but "draw this very specific picture that I have in mind" is specific enough).
Thanks for the link to the actual decision; the lawyer's summary included a link to download a PDF of the decision, but a direct link is useful to have.
... Actually, on double-checking, the lawyer's summary I linked to lets you download a PDF of the original judge's decision (written by Judge Beryl A. Howell). The link you provided is to the appelate court, affirming the original decision (appelate court opinion written by Circuit Judge Millett). So both links are useful.
As for your "Where did they say this?", that sentence (the amount of work will matter) is my summary. The original decision said the following (page 13 of the PDF of Howell's decision):
> The increased attenuation of human creativity from the actual generation of the final work will prompt challenging questions regarding how much human input is necessary to qualify the user of an AI system as an “author” of a generated work... This case, however, is not nearly so complex.
And then there's this from the appelate court decision, on pages 18-19 (italics in original):
> First, the human authorship requirement does not prohibit copyrighting work that was made by or with the assistance of artificial intelligence. The rule requires only that the author of that work be a human being—the person who created, operated, or used artificial intelligence—and not the machine itself. The Copyright Office, in fact, has allowed the registration of works made by human authors who use artificial intelligence. See Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Reg. 16,190, 16,192 (March 16, 2023) (Whether a work made with artificial intelligence is registerable depends “on the circumstances, particularly how the AI tool operates and how it was used to create the final work.”).
So although those words ("the amount of work put into the prompt will definitely matter") do not appear in either decision, the meaning is clear. Whether a work can be copyrighted will depend on unnamed circumstances including how the AI tool was used. The conclusion I draw from that is that "how the tool was used" is going to include "how much detail was specified", i.e. "draw a cat" doesn't make you the author, but "draw this very specific picture of a cat that I have in mind" will. That's my opinion, as I pointed out in the next paragraph — but now you know which parts of the decision(s) I based that opinion on.
> Whether a work can be copyrighted will depend on unnamed circumstances including how the AI tool was used. The conclusion I draw from that is that "how the tool was used" is going to include "how much detail was specified"
The Copyright Office's guidance the court cited said it would not.
If a work’s traditional elements of authorship were produced by a machine, the work lacks
human authorship and the Office will not register it 26 For example, when an AI technology
receives solely a prompt 27 from a human and produces complex written, visual, or musical
works in response, the “traditional elements of authorship” are determined and executed
by the technology—not the human user. Based on the Office’s understanding of the
generative AI technologies currently available, users do not exercise ultimate creative
control over how such systems interpret prompts and generate material. Instead, these
prompts function more like instructions to a commissioned artist—they identify what the
prompter wishes to have depicted, but the machine determines how those instructions are
implemented in its output.[1]
And the Copyright Office rejected a claim 624 iterations of experiment and refining was authorship.[2]
Another court could change this.
> That's my opinion, as I pointed out in the next paragraph
You said the next paragraph was your opinion in the next paragraph.
Right. It’s the right decision (for now) that only humans hold copyright.
There’s a whole other mess of fair rights and attribution but in this case, the ruling actually makes the path clearer for an author of such a piece to get it copyrighted. Ironic. But that’s how case law is supposed to be.
When I started my career it was in graphic design in the late 90s and photoshop was all the rage. There was a similar discussion as you pointed out around copyright and photo-manipulation technique. Who owns the photo? Who owns the piece? Then concept arts “kit bash” with reference photos to paint quickly… again who owns the refs? Who owns the piece? Deviantart had to struggle to figure out how to respect both.
I didn’t have to read too far to understand exactly where the court was going with this and where he could have just turned around with another filing and been gold.
> If you just copyright it normally with yourself as the author, seems like it would be fine to copyright whatever bs you want?
That's not really clear; the Copyright Office has guidance on copyright with AI used as a tool (which both complicated and lacks clear, objective standards), but its not at all clear how the courts would view it (I think there have been some trial court cases over it already, but I am not aware of any appellate law).
What is clear is that this case doesn't address that issue at all.
Photography is a copyrightable art, but don't say that most of the work is actually done by a machine. Or even by the engineer who built the machine. You could argue that the photographer just presses a button and, perhaps, airbrushes it later in photoshop, and yet that's art.
It's not copyrightable automatically, you have to argue that you did have an artistic input (e.g. composition). Typically nobody bothers to argue against copyrightability of a photo, but there's been a few cases.
Do you know of any sources that talk about this? I tried to do a bit of searching and the closest I found was the .gov site [0] that did make a similar-ish claim, but was vague enough (at least to me, a non-lawyer), that it doesn't seem to rule out that every photo taken by an individual is copywriteable
>First, copyright protects original works of authorship, including original photographs. A work is original if it is independently created and is sufficiently creative. Creativity in photography can be found in a variety of ways and reflect the photographer’s artistic choices like the angle and position of subject(s) in the photograph, lighting, and timing.
I find it hard to imagine a photo taken by someone where it couldn't be argued that those elements exist. I guess the photographer would have to explicitly tell the court something like "no, I put no thought into it whatsoever, the camera was hanging off my bag and the shutter button was pressed accidentally". Like, if a human purposefully took a photo, then they have made choices about location, subject, etc. which have some element of "creativity" to them.
It's a simple and quite recent Dutch case (feel free to use AI to translate it :p), where the courts basically said that the plaintiff did not sufficiently motivate why their photo would be copyrighted, especially in light of very similar photos having been made by other people (4.5).
Tell me you cant take good photos without telling me you cant take good photos. Photographers have the skill of colour, framing, perspective, and timing.... and if its nature, they also have to carry heavy bags of camera gear along with their hiking gear and bear spray etc and go out and then cpture photos....national geographic literally made its career off photos.....do you think about what orher weird takes you may have and never notice how wrong they are?
A common technique in paint print shops is to print the piece on canvas, then “add” color to it in globs of acrylic that match, making it stand off from the canvas. A very quick a clever trick to recreate multiples of a piece.
> 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. https://www.copyright.gov/newsnet/2025/1060.html
I think that this means that a single prompt alone does not convey copyright. But if you had spent many hours before the prompt fine tuning the model, or much effort after the prompt shaping the result with further prompts, it could be.
I disagree with this approach because I've seen how much creativity and effort some people can put into slowly evolving a single elaborate prompt. AI can be used as another kind of brush. A prompt can be a masterpiece.
I don't think this is the correct interpretation. I think they mean that if you make something without AI and then modify that with AI, that's covered. Likewise, if you start from an AI output and modify it, that's covered.
But the pure output of a generative model cannot be copyrighted, regardless of how complex the prompt is (note that the prompt itself could be copyrighted).
> But the pure output of a generative model cannot be copyrighted, regardless of how complex the prompt is
If that’s how the court interpreted it, then the software industry is hosed, since that’d mean none of the generated code running in production right now is under any sort of copyright or otherwise protection, lol.
I doubt that much software is entirely AI-generated with no human review or testing, it’s probably more like integrating some public domain snippets you found online into your code (which doesn’t invalidate copyright on the rest of it, or the way it’s put together) or having some files auto-generated by a script (like a C header containing a lookup table for a simple mathematical function, the table isn’t copyrightable itself maybe but the software as a whole still is)
If a deterministic machine transformation from a copyrightable prompt results in an uncopyrightable image, what do you think a compiler is doing to source code?
AI is not specifically not deterministic from the enduser's perspective. they throw randomness into it and hence why an exact prompt wont produce the same exact result.
a compiler on the other hand is generally pretty deterministic. The non determinism that we see in output is usually non determinism (such as generated dates) in the code that it consumes.
because they are just translating code (that everyone agrees is copyrightable) in a deterministic manner into another medium.
I'm not saying AI art should or shouldn't be copyrightable. One can argue the inputs into the AI generator are copyrightable, but if the output isn't deterministic translation of the input, its a different argument.
The original argument was that AI works wouldn't be copyrightable because they are deterministic, i.e. are just an algorithmic transformation lacking in creativity.
It sounds like they might be under the impression that having any AI-generated output in the code even if parts are human authored would invalidate the copyright, which isn’t true
>If that’s how the court interpreted it, then the software industry is hosed, since that’d mean none of the generated code running in production right now is under any sort of copyright or otherwise protection, lol.
I'm not sure this is really true, since copyright applies to distribution.
If you have a substantial amount of backend code (as with most SaaS projects) you're never actually distributing the code, and copyright is never at play. Computer generated artifacts are already in this boat and are protected by virtue of being trade secrets not by copyright.
This could maybe be true of shipping javascript to the browser, which presumably is not going to qualify as a trande secret, but I don't think that's where most companies derive value.
The idea that copyright applies solely to distribution is a popular myth, but it has no support in the actual copyright law. The core exclusive rights in copyright are (in the US, 17 USC § 106):
---
(1) to reproduce the copyrighted work in copies or phonorecords;
(2) to prepare derivative works based upon the copyrighted work;
(3) to distribute copies or phonorecords of the copyrighted work to the public by sale or other transfer of ownership, or by rental, lease, or lending;
(4) in the case of literary, musical, dramatic, and choreographic works, pantomimes, and motion pictures and other audiovisual works, to perform the copyrighted work publicly;
(5) in the case of literary, musical, dramatic, and choreographic works, pantomimes, and pictorial, graphic, or sculptural works, including the individual images of a motion picture or other audiovisual work, to display the copyrighted work publicly; and
(6) in the case of sound recordings, to perform the copyrighted work publicly by means of a digital audio transmission.
---
OTOH, distributing copies created in violation of copyright is a good way to cause legally-cognizable harms to the copyright holder that will increase the potential damage award when you are found liable for copyright infringement, and it also makes it much more likely that someone will notice the infringement in the first place. But its not where the law, on its own terms, begins to apply.
Doing any of those without permission (unless it falls into one of the exceptions to copyright protection, like fair use) is a violation of copyright.
The idea of copyright is to prohibit unauthorized use and reproduction, but none of this actually happens with a proprietary software SaaS backend. You don't actually give anybody the code - they connect to the service.
Access to the service is already governed by computer access laws, which don't depend on copyright. And if you never intentionally distributed your code outside of your org, you can call it a trade secret and nobody else has any legitimate right to access it - whether or not it is copyrightable.
There are other things that aren't copyrightable that are trade secrets already. This would be true of any kind of automated data collection for example. You couldn't copyright it but you can call it a trade secret.
And for any of that stuff, if you want to share it and limit distribution, you just have whoever wants access explicitly agree to be bound by contract law.
>The idea of copyright is to prohibit unauthorized use and reproduction, but none of this actually happens with a proprietary software SaaS backend. You don't actually give anybody the code - they connect to the service.
The point isn't that you have to give it to people, but okay?
>Access to the service is already governed by computer access laws, which don't depend on copyright
Yeah, copyright doesn't control everything, and?
>There are other things that aren't copyrightable that are trade secrets already. This would be true of any kind of automated data collection for example. You couldn't copyright it but you can call it a trade secret.
Okay?
>And for any of that stuff, if you want to share it and limit distribution, you just have whoever wants access explicitly agree to be bound by contract law.
Your point being? You're just rambling assumptions about copyright and other things, which don't even track the actual law.
> Your point being? You're just rambling assumptions about copyright and other things, which don't even track the actual law.
I'm replying to the post that claimed:
> If that’s how the court interpreted it, then the software industry is hosed, since that’d mean none of the generated code running in production right now is under any sort of copyright or otherwise protection, lol.
There is in fact "otherwise protection" for the software industry by... not distributing the code. They don't need copyright over the generated code if they vibe code a SaaS backend. Whether there's copyright or not is irrelevant for the business model.
Copyright is the strongest legal protection available. It does not have a state of mind element. Breach of contract is much more complicated and context-dependent.
>There is in fact "otherwise protection" for the software industry by... not distributing the code.
Copyright protects against reverse engineering in some circumstances, for example.
>Whether there's copyright or not is irrelevant for the business model.
Yeah, I'm going to continue to disagree with you as I'm actually a litigator.
> Yeah, I'm going to continue to disagree with you as I'm actually a litigator.
OK, can you explain to me why this is a disaster for a vibe-coded SaaS? Why are computer access and/or contract laws insufficient and why would a vibe-coded backend be a huge risk?
I really don't understand where copyright on the code itself is necessary to protect these business models, and hopefully you can help fill the gaps.
I didn't say it would be a huge risk, I just disagree that any of those features of the law cover what copyright does. They don't. If a trade secret is ever revealed, all protection is lost. Breach of contract is very complex compared to an infringement claim and would have to be negotiated. As a customer, why would I want to indemnify a software supplier? If there's no indemnity, it's not going to get anyone very far. CFAA basically requires that something get hacked so it's not going to cover the vast majority of scenarios...
>I really don't understand where copyright on the code itself is necessary to protect these business models, and hopefully you can help fill the gaps.
Well, did you ever try to understand? It's so exhausting coming to these threads when people are just making assumptions about how the law works without any regards to what actually happens, and then suggesting policy changes in response.
Here's a scenario - disgruntled ex-employee leaks the code. Now it's free for anyone to use because there is nothing you can do to stop anyone from using it because you have no rights in the code since the trade secret is broken. You can sue the employee. They are probably judgment proof, wont have a lot of money anyway, and will still not stop a competitor from spinning up the same exact thing.
Trade secret was your suggestion by the way... So do, you actually know how trade secrets work, or you just making things up??
What constitutes a modification? Here’s a reverse Sorities Paradox situation.
Let’s say I use an AI prompt to generate an image with 24-bit color, and then I manually change the RGB value of a single pixel from (255,255,255) to (254,255,255).
Does that constitute a modification and would then allow the image to be copyrighted? If not, where is the line?
This is how I understood the original decision a while back - that there had to be some additional element of human involvement post-"gen", though to what extent is still a bit unclear to me.
What's the threshold? Can the person just slap an LUT on an SDXL image in Photoshop and call it a day?
Its interesting you mention a toddler drawing for 2 hours - I can say with certainty that toddler drawing will demonstrate more creativity than this argument and, every iteration of it, I see every single day. Smh
Actual creators understand what creativity is and what creation is - not all creation even is creative, as its really more of a process, than it is a singular output, and there are monotonous, time consuming, meticulously frustrating parts of the process of creation.
If you want to limit yourself to creation without the quality of life enhancements, that exist in this time and space - you do you.
Don't proselytize tho - you are not doing what you think you are.
> If some toddler spends two hours on a drawing, it's a masterpiece?
A work doesn't have to be a masterpiece to obtain a copyright so what difference does it make? The point of copyright isn't to subjectively critique art.
The output is mechanically derived from the prompt. There's zero artistry or creativity involved, and despite all the posturing, those are more or less hard requirements for copyright. The prompt may meet the criteria and be copyrighted.
The outputs of a camera 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 photographic image, or a human makes creative arrangements or modifications of the image, but not the mere pointing of a lens and adjustment of setting.
I don't think that word means what you think it means.
You have an extremely low bar for calling something a masterpiece.
A prompt can be clever, insightful, unique, and even uniquely productive.
But it is nowhere near the level of decades-deep skill and creative inspiration required to create art anything worthy of the label "masterpiece".
>>AI can be used as another kind of brush
Perhaps that is a valid analogy, but we do not give copyrights to brushes, no matter how much cost or effort was required to make the brush. The brush is not the only tool required to make the art. To continue the analogy, the artist must also select and mount the canvas, mix and color each shade of paint, build up the base layers, and on and on and on...
It doesn't matter if your "brush" is a five hundred billion dollar machine and you spend six months whispering to it to find just the right incantation to generate your file of pixels — SCOTUS is right, you have not make art to which you can claim a copyright.
And the starving student artist in their garage mixing their paints and using the dollar-store brush did make art worthy of a copyright claim.
>> I disagree with this approach because I've seen how much creativity and effort some people can put into slowly evolving a single elaborate prompt. AI can be used as another kind of brush. A prompt can be a masterpiece.
Absolute nonsense. A work of art is made of semantic stratification, experience, thought process.
A prompt lacks all that.
AI art can be a tool, but this sentence is a good reminder that on average it’s worth shit all.
Don't forget that the human artist spends hours training on art that predates them, and, in my opinion, that training predisposes the artist to unconsciously replicate elements of art they've trained on previously.
Am I treating humans like machines or machines like they're human?
LLMs were created in our image. Hallucination, confabulation, sycophancy, psychopathy, learning, reasoning, and blackmail are all behaviours in LLMs that were first found in humans. All these behaviors are present in human writing and imagery captured in a training set. So to me, there's no surprise that LLMs exhibit these behaviors.
Do I think LLMs are sentient or sapient? I'm in the probably not camp. We don't have a good test for either, but they do illustrate the resistance to acknowledging any other being or creation as having the same capabilities as Homo sapiens.
> AI can be used as another kind of brush. A prompt can be a masterpiece
What a joke. No, AI is not a brush, it is a slop machine that spits out derivatives of the actual masters. If you go back and forth with a human artist about a commission where you keep nitpicking and wanting adjustments, does that make you the artist? No, it makes you the “ideas guy”
As someone who has actually studied art - right back at you.
All the most famous paintings are done like a hundred times - certain scenes were done by ALL Masters, as like a rite of passage or proof of arrival - pretty much the whole time between the OG Masters and modern art.
"Who is your favorite Sebastian?" Is a legit question, for example.
And? Have we all already lost any ability for actual consequential thinking?
Each one of those Sebastian is the living proof of a human experience, each with its subtleties, each unique. That is why you ask that question.
How on Earth can you compare that to a series of shitty pictures done in a few seconds by a stochastic machine?
You clearly only get sloppy from the machine -> hence your entire understanding, you did some googling, found others with the same experience and you took up this position.
Meanwhile, random person, gets the exact same AI that you used to create literal DaVinci'esque, visibly masterpiece inspired - maybe not "masterpiece" but "masterpiece adjacent" - thats apparently, its not perfect art, but it could have been created in a workshop...
You can't do that. Rather, you cant nake the AI do that.
What is the difference between you and the random person with artworks in the style of the old masters? What do we call that gap?
Isn't that gap normally stuff like talent, ability, skill, knowledge?
All arguments made in this vein are just people whining about their personal lack of ability, as if its a machines fault.
Thanks for the ad hominem. Subjective artistic value aside, do you not dispute that prompting AI to create an artistic image is functionally the same as outsourcing to another human? You input instructions and a commission fee; you get back a piece of artistic expression. You did not create the art, someone else did. And for AI output, the machine cannot claim a copyright on that original image.
If you outsource the image creation, then throw it on a t-shirt design, or modify it, or do whatever, you can copyright the modified work that you Han a hand in creating, but you still are not the creator of the original.
Let's say the AI prompt is "Make it black and white". Why does taking a photo and making it gray scale in photoshop result in a copyrightable piece of art, but using an AI model makes the resulting output slop? They seem equivalent to me.
I'll distill it down into something you might understand a bit easier. On social media, such as Instagram, or Tiktok, you'll find a bit of a meme going around that shows the difference between an influencer video of a vacation destination and then a follow-up video from someone with their iPhone, often showing overcrowding of tourists, brown water where there was blue; with these videos often with a poor-sounding Recorder being played over them (I forget the song that it's attempting to play).
The difference between the "real" video and the "influencer" video is the artistry from the artist(s) involved. (And yes, top influencers often have a person or a team of people involved)
Because you're unable to understand art, honestly. Photography isn't just "taking a photo and making a grayscale image in Photoshop"; but rather a combination of a couple of different artistic expression styles that involve understanding how to use the tools you have (a camera, the lens, film or a sensor, and lighting) to capture an expression of an event. Technically speaking, a photo of a mountain isn't just a "photo of a mountain" that you would maybe throw into AI slop--but an actual, legitimate photo of a mountain and how it's captured and presented (no matter the post processing done) is an actual artistic expression of the capture of that mountain. Because absolutely nothing, nothing at all will capture actually standing there looking at the mountain. A photo is the best approximation, and sometimes doesn't even have to be, depending on what the artist wants to express with the image.
In short, your inability to understand photography doesn't justify the use of AI slop to prompt "give me a grayscale image of a mountain" and assume that it's the same thing as a human being taking an actual photo. They're not even close to the same thing.
They are describing the artistic qualities that something must posess to be actual artwork - which is relevant to copyrighting artworks, and is also that had been discussed, by Artists and Creators - the entire time
They did not write an emotional comment, they were trying to teach you, bc they ascertained that you don't understand art.
>use of AI slop to prompt "give me a grayscale image of a mountain"
That is not the prompt I included in my post. The prompt I gave was for taking an the same photo you would be putting into photoshop and using AI to apply grayscale.
Initialize an algorithm to point your camera at the street and write those bytes to disk and you are the author of a perpetual stream of data.
Initialize an algorithm to point your camera at the street and describe those bytes in words and you are no longer the author a perpetual stream of data.
In a world where slapping an overlay of someone looking incredulous over someone else's video is considered "adding substantive commentary" by every major video sharing platform, I don't even try to understand copyright law at all. It is way over my head.
If it makes seven figures of revenue, there is a real system in place to litigate copyright disputes between corporations. Two kaijiu summoned by ritual magic to fight for the future of the franchise / giant pile of money.
Everything else in the entire system is just bits of monster and building falling randomly. We know if we put the whole population under strict scrutiny ("laser eyes" + "lightning wings"), it would kill every last one of them; every teenager is theoretically criminally liable for the GDP of the Milky Way, a series of violations beginning with a performance of The Birthday Song at their first cake day. Even hiring the cheapest defense lawyer would bankrupt nearly any family in the nation. So we try imperfectly to dodge copyright, hopefully by a couple zip codes, and live in a state of nature on the ground.
> slapping an overlay of someone looking incredulous over someone else's video is considered
it really isn't, you actually have to provide enough relevant commentary for it to be transformative
it just looks like that because
- not every claim leads to a take down, more common is that the advertisement revenue is redirected to the owner of the original video. That is very very common, especially on YT, but not really visible as viewer.
- there are enough copyright holders which overall tolerate reactions, even if they don't fall under fair use.
- Sometimes people claim it doesn't fall under fair use when they don't like how the reaction is done, but that doesn't mean it wouldn't be ruled fair uses if it came in front of court.
- Sometimes people reacting have explicit permission from the original author to do so, no matter if it counts as fair use or not.
and maybe most relevant here, pretty much all large platforms have a tendency to favor the person claiming the copyright violation over the person which reacted to it. To a point there is is sometimes a big problem if systematically abused with false claims.
Those two hypothetical scenarios you listed don’t necessarily work the way you are describing it, which is why the whole logic and mechanisms behind the US copyright laws might seem incomprehensible or illogical to you.
In reality, it is way more complex and less clear-cut. Which makes sense, because oversimplifying it will lead to silly-sounding conclusions and an almost entirely incorrect understanding of how this works.
For those who don’t want to read the actual full explanation (which is a totally normal position, as the explanation is going fairly into the weeds), I will just a put a TLDR summary at the end. I suggest everyone to check out that summary first, and then come back here if there is interest in a more detailed explanation.
----------------------------
First, we gotta settle on 3 key concepts (among many) the US copyright law relies on.
1. Human authorship - self-explanatory; you cannot assign authorship to a fish or your smartphone.
2. Original/minimal creativity - some creative choices, not just "I pressed the button."
3. Fixation - the content needs to be recorded on a tangible medium; you cannot copyright a "mood" or a thought, since those aren’t tangible media.
Now onto your hypothetical scenarios:
1) "Initialize an algorithm to point your camera at the street and write those bytes to disk and you are the author of a perpetual stream of data."
Writing bytes to disk satisfies fixation, but it doesn’t automatically make you the author of a copyrightable work. You gotta satisfy the minimum creativity requirement too (e.g., camera positioning, setup, any other creative choices/actions, etc.). Otherwise you are just running a fully automated security cam feed with zero human input, and those videos aren’t easily copyrightable (if at all). You might own copyright in a video work if there’s sufficient human creative authorship - but mere automated recording doesn’t guarantee that.
2) "Initialize an algorithm to point your camera at the street and describe those bytes in words and you are no longer the author a perpetual stream of data."
This is just close to being plainly incorrect. If you (a human) write a textual description, that text is typically copyrightable as a literary work (assuming it’s not purely mechanical like "frame 1: car, frame 2: another car, etc." with no expressive choices). Creating a description doesn’t erase any copyright you may or may not have had in the underlying recording. They’re just different works (audiovisual work vs. text work).
Important to note: neither makes you the author or owner of the underlying "data" of reality, because copyright protects expression, not the underlying facts.
----------------------------
TLDR:
* Recording the street can produce a copyrightable work if there is human authorship and minimal creativity in how the recording is made. Pure automated capture may fail that.
* Describing the street in words is usually a separate, independently copyrightable work (e.g., a text or audio version of those words), but it doesn’t change the status of the underlying recording.
But how does that apply to photography vs AI photo generation?
Photo (w/ camera):
1. MET: Human authorship - somebody picked the tools (lens, body) and used them.
2. MET: Creativity - somebody chose a subject, lighting, etc.
3. MET: Fixation - film (or SD card)
Photo (w/ AI):
1. MET: Human authorship - somebody picked the tools (models etc) and used them.
2. MET, maybe?: Creativity - somebody wrote the prompt, provided inputs, etc. (how is this substantially different than my wife taking a random snapshot on her phone?)
3. MET: Written to disk, same as a digital camera.
The camera analogy breaks at one specific point: who determines the expressive elements of the final work.
With photography, the human determines framing, angle, timing, lens, exposure. The camera just records light from a scene the human selected and composed. Even a random photo reflects where the photographer stood and when they pressed the shutter. The device doesn’t invent the composition.
With AI imagen, the user provides high-level instructions, but the system determines the actual composition, lighting, geometry, textures, etc. The expressive details of the final image are generated by the model, not directly controlled by the user.
That’s why the US copyright laws currently treat them differently. It is less of a "tool vs. tool", and more of whether the human determined the expressive content (or if the system did). Prompting can be creative (in a legal sense), but giving instructions is not the same as controlling the expression.
If I tell a human painter “paint XYZ in an expressionist style,” I don’t become the author of the painting. The painter does, because they determined the expression. And since the painter (in the case of AI imagen) is not a human, then that work usually cannot be copyrighted.
There is an important caveat to all of this: it’s not binary or perfectly clear-cut. If someone iteratively refines prompts, controls seeds, manually inpaints, selects and arranges outputs, heavily edits the result, etc., then those human contributions can be protected. But purely AI-generated output, where the system determines the expressive elements, is not considered human-authored under the current US copyright laws.
Mind you, none of this is perfectly settled, as this is a very rapidly evolving/adapting area of law (as it pertains to AI usage). I am not claiming that this is the end-all of how it should be legislated or that there are no ways to improve it. But the current reasoning within the US copyright law used to address this type of a scenario (at the present moment) doesn't strike me as illogical or unreasonable.
AI-generated art can't be copyrighted, fine. But what does this mean for the huge spectrum between "I did some fingerpainting" and "Nano Banana spat out this painting"?
What if I use Photoshop and context-aware fill a cloud in? Is that AI-generated or human-generated art?
most likely counts as AI-assisted art, which is copyrightable with you as the owner
like most things copyright there is a gray area there
but in most cases it's either pretty clear and courts would most likely rule in your (copyright holder) if you somehow manage to hit the perfect middle of the gray area
through if you tell the court "the author is my AI" (like in this case), the outcome is pretty obvious
also for better understanding using AI doesn't erase copyright, it just doesn't add it. So if you image was copyrightable before you used an AI tool to change it will stay copyrightable (as long as the original image is still in there to a reasonable degree).
I wonder if Nano Banana spits out an image and I copy it by hand into a different medium like acrylics, chalk, or charcoal. Does a manual transcription suddenly render the image worthy of copyright?
This also raises the meta question: how much does an image need to change to acquire a new copyright? For example, if you change the Last Supper to include two fat Jesuses on either side of the single skinny Jesus, is that enough?
A photograph or a tracing made with an optical device is a mechanical reproduction of the original work, but it has its own copyright because there's a human in the loop.
A forger takes skill and creativity to make a copy of an existing work that is so good it cannot be distinguished from the original. By your assertion, the forgery should have its own independent copyright.
For many, many reasons, I suspect patents will become much more important now. After all, it's the ideas that matter now. Which I maintain, has always been the case, because "execution" is nothing more than a series of smaller ideas, except those typically needed money. How convenient for those with capital!
Patents have the drawback of being expensive and very slow to acquire, but having worked on a bunch, they are uniquely suited to be radically optimized by GenAI.
Also patents are very flawed in practice, but the only real protection that is left. Copyright is meaningless when, as people have done, you can reproduce entire saas products by feeding AI screenshots.
Intellectual Property as a whole has been in need for a revamp for a while now, but it's even more critical in the age of AI.
Patents are what allowed the industrial revolution to happen. No one is bringing a cotton gin from idea to design to manufacturing to market if the second you release it every manufacturing company in the world can start making their won.
This strawman pops up often. The purpose of IP is to "promote" progress. The word was carefully chosen.
And luckily we have years of research, empirical and otherwise, showing the impact of IP on innovation and the arts. It's a very complex, multi-dimensional topic where outcomes vary depending on a combination of things like industry, subject matter, time frame, level of economic development of the country, strictness of enforcement, and more.
However overall the impact of IP is much more positive than HN would think, largely because the material and discourse that has gotten airtime here is something that aligns more with the "information wants to be free" crowd.
People are taking screenshots of existing products and feeding them to AI to reproduce them from scratch. From this site: https://ghuntley.com/real/
> Any product features or platforms that were designed for humans. I know that's going to sound really wild, but understand these days I go window-shopping on SaaS companies' websites for product features, rip a screenshot into Claude Code, and it rebuilds that product feature/platform.
> Patents aren't going to cover the vast majority of what exists as code because patents only cover inventions and not common every day code.
Agreed, but as we're finding out, everyday code is now cheap enough that there will be questions about how much of it is worth protecting... and, given the unclear stance on AI-generated code, maybe even if it is protectable at all.
Yes, the bar for patents is way too high for every day code, but you can always get one by making the claims narrow enough. I think that will actually serve one of the underlying goals of patents by incentivizing people to build things that are actually novel and non-obvious rather than just a slight variation of another project.
For projects where there is absolutely nothing technically novel (which is rarely true, cf "claims narrow enough") moats could be in hoarding data, network effects, and the like.
Note, I'm not necessarily happy about all this, this is just how I think it could play out absent larger changes to IP laws.
>Agreed, but as we're finding out, everyday code is now cheap enough that there will be questions about how much of it is worth protecting... and, given the unclear stance on AI-generated code, maybe even if it is protectable at all.
AI generated code is likely not protectable at all under copyright, but that was expected by anyone who follows this area of the law seriously.
> Yes, the bar for patents is way too high for every day code, but you can always get one by making the claims narrow enough.
Not really, and if you did it would likely not be enforceable because the getting past a §101 motion isn't likely.
>Note, I'm not necessarily happy about all this, this is just how I think it could play out absent larger changes to IP laws.
I'm not sure any of it is a bad thing. Why should AI generated code be protectable under copyright? It can still be protected by trade secret, which seems fairly sufficient for what it ultimately is.
Yes "deduced" makes more sense, though I guess the code was not really the point. I really should have said the backend functionality is being reproduced. As copyright doesn't protect functionality, the closest proxy we had was the code itself, and now that is meaningless.
> AI generated code is likely not protectable at all under copyright, but that was expected by anyone who follows this area of the law seriously.
I haven't followed too closely, but the ruling on AI art seems to leave the door open for "significant human creative input" like "editing, arrangement." This leaves room for photography, but can also be covered by "architecture, followed by iterative prompting and potentially manual editing to refine the code", which is still needed for any non-trivial AI-generated code.
I think the situation is not great because the law is too fuzzy and outdated, and could allow for "AI laundering" of copyrighted works, as the relicensing thread suggests, which is especially not great for open source where trade secrets are not an option.
> Not really, and if you did it would likely not be enforceable because the getting past a §101 motion isn't likely.
Not a lawyer, but I've seen tons of such issued patents out there, and many have even been upheld in lawsuits. §101 is often trivially bypassed adding "computer readable media" or some "non-abstract side-effect" language to claims. Like, one pattern of claims is "A method of - algorithm - algorithm - algorithm - show something to the user / write to a DB / transmit a message / etc."
Perhaps. Even that may not be important if the METR progress line continues much longer, because then all those billion dollars "worth" of software written over the past 3-ish years get re-invented for cents on the dollar.
Separately, I think code is more like an invention than a work of art, and should have been subject only to patent laws instead of (and not in addition to!) copyright laws. This doesn't really make much difference now, as AI doesn't (at least in the UK) have personhood for either copyright or patent law: https://www.briffa.com/blog/can-you-obtain-a-patent-for-inve...
I would assume that the same
“substantial human authorship” criterion applies. Copyright is about human creativity, it doesn’t otherwise matter if something is art, prose, typesetting, or code.
> The Supreme Court previously rejected Thaler's request to hear his argument in a separate case involving prototypes for a beverage holder and a light beacon concerning whether AI-generated inventions should be eligible for U.S. patent protection. His patent applications were rejected by the U.S. Patent and Trademark Office on similar grounds.
That would effectively rely on the doctrine of trade secret rather than copyright. A major difference is that accidental or malicious disclosure of a trade secret usually ends the trade secret status, forever. In an alternate universe where computer source code had never been copyrightable, famous leaks (Microsoft Windows, 2004; id Quake, 1997) would have effectively open-sourced those codebases, and other companies could have openly and legally used them.
As source code becomes more of a generated artifact of software development the way object code is an artifact of compilation, we might be moving toward a world where secrecy, constant forward motion, and moats become even more of an asset (vs plain IP protection).
I do think what happens in this case is SCOTUS will ultimately rule that AI-built code is copyrightable while art is not. I'm sure there's some rationale thick enough for them.
It's strange how hard it is to think of a situation that could lead to that case. Who would bother filing an infringement lawsuit for code whose very existence proves that it can be derived by anyone from LLM prompts? What would the damages even be?
Interesting world we live in. Soon it'll be faster to one-shot the tiny slice of functionality I need from Adobe CS than to navigate their subscription cancellation obstacle course.
> Soon it'll be faster to one-shot the tiny slice of functionality I need from Adobe CS than to navigate their subscription cancellation obstacle course.
This is precisely why copyright is practically obsolete. You can't legally forbid someone from paraphrasing, and now we can easily automate it to just within the threshold set by legal cases.
I view this as a massive personal challenge. Can I write instructional materials that are better than an AI summary? I'm going to keep trying, even as books become obsolete.
Can you imagine the chaos if suddenly all the slop code wasn’t owned by the company? Even though that result would be consistent with this ruling, it undermines the narrative the economy is now riding on, so there will likely be special exemption.
I don't think this article's analysis is accurate. The "human authorship" in Thaler's case wasn't about the abstract concept of human authorship; he literally did not put his name in the "authorship" field of the form, and insisted on review that his name doesn't belong there because he's not the author.
So the ruling doesn't necessarily endorse the Copyright Office's analysis referenced in the article (https://www.copyright.gov/ai/Copyright-and-Artificial-Intell...), and I think that analysis is just not correct. They describe a creator of AI art as simply "re-rolling the dice" when they try different prompts, but that's not correct, clever prompt engineering definitely allows you to "constrain or channel the program’s processing of the sourcenmaterial" and "alter[] the degree of control over the process"
Long standing well known issue, no copyright (in many countries) and (in some countries) non patentable, too.
Through this isn't true for AI assisted art.
And the gray area is very wide and very legal unclear (gray area between human art with AI assistance (e.g. "AI"/transformer architecture based line smoothing or color calibration) and AI art with human touch added to it).
Note that this has very little bearing on the real interesting questions of whether and when human authors can copyright works where AI was used as a tool; this case is specifically about attempts by Thaler to apply for copyright listing an AI as author of a work for which he explicitly denied any human authorship.
Exactly 0 of the artists I know “generate” their images outright - AI generations are always part of a pretty typical artistic workflow.
The way I think of it is this: typical art creation starts from a blank canvas and the artist adds layer upon layer of what you want. Eventually something coherent (to the artist at least) pops out.
AI art starts from a canvas which is filled, and the artist changes the filled canvas to meet their perspective. It’s like those projects where people take a vintage painting and add Pokémon to it. Mostly the people I see using AI art are traditional artists who view it as a new medium in their process, very few “generate” and call it a day.
If we're going to allow AI companies to use copyrighted material in training, the absolute least we could do is prevent copyright of the outputted content.
Similarly, right now AI art is widely dismissed as "just prompts." But having tried many times to generate images via prompts, it's very hard to get what's in my head to show up in the result. I ended up spending much more time editing the images than creating them... but, I could do that with much simpler tools, without learning advanced tools like Photoshop.
In a couple of instances though, the AI has blown me away by generating something that better captured what I wanted to convey! I suspect the trick is in beng very detailed in where I was coming from and the emotions I wanted to engender.
I predict appreciation of AI art will shift to overall imagination, taste, and appreciation of technical nuances noticeable only to those "skilled in the art", such as prompting techniques and the quirks of the model used. I even suspect there will be genres of AI art using weaker models (kind of like photographs with Polaroid cameras.)
The main reason this is getting really complicated is because copyright inherently goes against the flow of how nature works. It's irrational at its core. It requires increasing amounts of maintenance to sustain.
It doesn't even promote the progress of science and useful arts anymore, and in fact when mixed with our current form of capitalism, it hinders it.
AI usage doesn't remove copyright, it just doesn't gain any new copyright by itself
so a AI based transformation of a copyrighted input is as much a potential copyright violation as a non AI based transformation.
It's just that the human transformation can by potentially itself be seen as art, so if you have a license or fair use you now can have copyright on the transformed peace (with some limitations (1)). And if the transformation is done by AI you won't (but the original authors "partial" copyright on the outcome is still there).
(1): Like if you (human) "transform" a peace of art in 1000 different ways each keeping 0.001% of the original you will likely get 1000x copyright. But if you then use this 1000 peaces you have copyright too to regenerate the original you still have full copyright infringement. In general the law doesn't care about your "trickery" trying to bend laws.
Every time I open my phone and find myself back on this comment thread, I find new nonsense.
If you are "anti-AI" and you’ve never changed or evolved your argument - I suggest a pause, a step back and a substantial revaluation.
Some of these comments in this thread - have me wondering if they have actually interacted with an AI.
You are not correct on "principle" - this isn't a moral thing, if you have taken an ethical position - its bc you dont have a functional understanding of how to make it function.
If you were functionally interacting with AI, you would have a more substantial postion, with actual criticism that would have value.
I'm reading a lot of sloppy- written by people, about AI slop.
Why is "AI-generated code" not also "AI art"? What makes "AI-generated code" copyrightable then? Nothing! Being that everything will be made using AI in the future, the courts just suicided the copyright system! Or where exactly does art end and code begin? The same applies to documents and designs.
If I take your AI-generated code file and write it as an artsy-looking image, do I get to deny you copyright?
No-one has checked in court of AI-generated code is copyrightable.
Personally I hope it's not. To me, this is the best outcome for AI in general. If we are going to violate everyone's copyright training AIs, then it's only fair you don't get AI protection on the output.
It is, and if you did what the actor in this case did with code (tried to register a copyright on a piece of code disclaiming any human authorship and claiming an AI tool as the sole author) it would also not be copyrightable.
This case does not actually address what people are actually mostly interested in with copyrightability of works made by humans with the assistance of AI tools, where the human running the tool is claiming authorship, because that was not what happened that was being fought over here.
so yes it applies to fully AI-generated code as much as to "AI art".
like with AI assisted art it doesn't apply to AI assisted coding
and yes if everything is fully AI generate there is no copyright anymore, that is by design!
Copyright is there to protect human creativity/time investment. If there is no creativity/time investment, then there is no reason for copyright to exist either. Having still copyright there would mean moving it from a law to protect creative work to a law to protect the privileged few which can afford to just mass generate "everything" with AI. That isn't just very undesirable, it's kinda plain evil, as it would mean screwing over the majority of humanity.
Naturally as mentioned that only applies to full AI products, not to AI assisted products in which case the "human contribution" and thinks resulting from it still have copyright.
> Thaler’s request to copyright an image, called A Recent Entrance to Paradise, on behalf of an algorithm he created.
The courts just take issue with him naming his AI system as the sole author and himself as the copyright owner.
If you just copyright it normally with yourself as the author, seems like it would be fine to copyright whatever bs you want?
No. The courts said what he could have copyrighted was a complex question they ignored because he said he was not the author.
Bingo. If he was the creator and owner, it would have been fine.
Not necessarily. The referenced guidance [0] says: "...copyright will only protect the human-authored aspects of the work, which are 'independent of' and do 'not affect' the copyright status of the AI-generated material itself." If you read the paragraph or two above that one, it really seems like products of agentic coding cannot be copyrighted, as there wouldn't be significant authorship involved.
[0]: https://www.federalregister.gov/d/2023-05321/page-16193
Very glad someone actually read the decision and understood it, despite how much reporting on this has been poor. This was not a case about "can AI-generated art be copyrighted?", despite all the reporting misleading people. (Including me, until somebody finally pointed me at the actual decision — https://www.lexology.com/library/detail.aspx?g=24e0581c-2c28... — and I could read it for myself). The judge literally quoted that case where a monkey picked up a photographer's camera and triggered it, saying that only humans can hold copyright: not animals, and not tools. And he also specifically said that he was not addressing "how much input is necessary to qualify the user of an AI system as an ‘author’ of a generated work".
So it's not the case, contrary to what many people (including me!) have said before, that the decision was "works produced by AI tools cannot be copyrighted". Rather, it's "you cannot assert that the AI tool itself is the author, you must assert that a human is the author". And the amount of work put into the prompt will definitely matter.
In other words, if you just prompt “draw a picture of a cat” then it’s possible you didn’t put enough work into the image to count as the author. But if you have a specific picture in mind that you want to create, and you prompt “draw a picture of a two-year-old cat with orange fur and orange eyes, in a sitting position, looking out of the window of a train. The interior of the train is lit with dim orange lighting. Outside, it is night and there is a full moon visible through the train window,” and then you refine that prompt until the AI produces an image close enough to what you had in your mind’s eye, then that image is clearly your own creation: the AI tool was just the tool you used to take the idea in your head and turn it into an image that other people could look at. Whether you use a paintbrush, a digital-art creation tool like Krita, or a digital-art creation tool like Midjourney, as long as you came up with the concept and did the necessary work to make the tool produce the image, then you're the author and you can assert copyright. (Note that this paragraph is my own opinion, not the judge's ruling, but I think it's a pretty defensible opinion: "draw a picture of a cat" might not be specific enough to assert that you created the resulting image, but "draw this very specific picture that I have in mind" is specific enough).
You pointed to what lawyers wrote about the decision. Not the decision.[1]
> And the amount of work put into the prompt will definitely matter.
Where did they say this?
[1] https://law.justia.com/cases/federal/appellate-courts/cadc/2...
Thanks for the link to the actual decision; the lawyer's summary included a link to download a PDF of the decision, but a direct link is useful to have.
... Actually, on double-checking, the lawyer's summary I linked to lets you download a PDF of the original judge's decision (written by Judge Beryl A. Howell). The link you provided is to the appelate court, affirming the original decision (appelate court opinion written by Circuit Judge Millett). So both links are useful.
As for your "Where did they say this?", that sentence (the amount of work will matter) is my summary. The original decision said the following (page 13 of the PDF of Howell's decision):
> The increased attenuation of human creativity from the actual generation of the final work will prompt challenging questions regarding how much human input is necessary to qualify the user of an AI system as an “author” of a generated work... This case, however, is not nearly so complex.
And then there's this from the appelate court decision, on pages 18-19 (italics in original):
> First, the human authorship requirement does not prohibit copyrighting work that was made by or with the assistance of artificial intelligence. The rule requires only that the author of that work be a human being—the person who created, operated, or used artificial intelligence—and not the machine itself. The Copyright Office, in fact, has allowed the registration of works made by human authors who use artificial intelligence. See Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Reg. 16,190, 16,192 (March 16, 2023) (Whether a work made with artificial intelligence is registerable depends “on the circumstances, particularly how the AI tool operates and how it was used to create the final work.”).
So although those words ("the amount of work put into the prompt will definitely matter") do not appear in either decision, the meaning is clear. Whether a work can be copyrighted will depend on unnamed circumstances including how the AI tool was used. The conclusion I draw from that is that "how the tool was used" is going to include "how much detail was specified", i.e. "draw a cat" doesn't make you the author, but "draw this very specific picture of a cat that I have in mind" will. That's my opinion, as I pointed out in the next paragraph — but now you know which parts of the decision(s) I based that opinion on.
> Whether a work can be copyrighted will depend on unnamed circumstances including how the AI tool was used. The conclusion I draw from that is that "how the tool was used" is going to include "how much detail was specified"
The Copyright Office's guidance the court cited said it would not.
If a work’s traditional elements of authorship were produced by a machine, the work lacks human authorship and the Office will not register it 26 For example, when an AI technology receives solely a prompt 27 from a human and produces complex written, visual, or musical works in response, the “traditional elements of authorship” are determined and executed by the technology—not the human user. Based on the Office’s understanding of the generative AI technologies currently available, users do not exercise ultimate creative control over how such systems interpret prompts and generate material. Instead, these prompts function more like instructions to a commissioned artist—they identify what the prompter wishes to have depicted, but the machine determines how those instructions are implemented in its output.[1]
And the Copyright Office rejected a claim 624 iterations of experiment and refining was authorship.[2]
Another court could change this.
> That's my opinion, as I pointed out in the next paragraph
You said the next paragraph was your opinion in the next paragraph.
[1] https://www.copyright.gov/ai/ai_policy_guidance.pdf
[2] https://www.copyright.gov/rulings-filings/review-board/docs/...
Right. It’s the right decision (for now) that only humans hold copyright.
There’s a whole other mess of fair rights and attribution but in this case, the ruling actually makes the path clearer for an author of such a piece to get it copyrighted. Ironic. But that’s how case law is supposed to be.
When I started my career it was in graphic design in the late 90s and photoshop was all the rage. There was a similar discussion as you pointed out around copyright and photo-manipulation technique. Who owns the photo? Who owns the piece? Then concept arts “kit bash” with reference photos to paint quickly… again who owns the refs? Who owns the piece? Deviantart had to struggle to figure out how to respect both.
I didn’t have to read too far to understand exactly where the court was going with this and where he could have just turned around with another filing and been gold.
Nah its theft
> If you just copyright it normally with yourself as the author, seems like it would be fine to copyright whatever bs you want?
That's not really clear; the Copyright Office has guidance on copyright with AI used as a tool (which both complicated and lacks clear, objective standards), but its not at all clear how the courts would view it (I think there have been some trial court cases over it already, but I am not aware of any appellate law).
What is clear is that this case doesn't address that issue at all.
Claiming that you digitally painted it? (Fraud?)
I mean, I guess lying is something you could do.
Photography is a copyrightable art, but don't say that most of the work is actually done by a machine. Or even by the engineer who built the machine. You could argue that the photographer just presses a button and, perhaps, airbrushes it later in photoshop, and yet that's art.
It's not copyrightable automatically, you have to argue that you did have an artistic input (e.g. composition). Typically nobody bothers to argue against copyrightability of a photo, but there's been a few cases.
Do you know of any sources that talk about this? I tried to do a bit of searching and the closest I found was the .gov site [0] that did make a similar-ish claim, but was vague enough (at least to me, a non-lawyer), that it doesn't seem to rule out that every photo taken by an individual is copywriteable
>First, copyright protects original works of authorship, including original photographs. A work is original if it is independently created and is sufficiently creative. Creativity in photography can be found in a variety of ways and reflect the photographer’s artistic choices like the angle and position of subject(s) in the photograph, lighting, and timing.
I find it hard to imagine a photo taken by someone where it couldn't be argued that those elements exist. I guess the photographer would have to explicitly tell the court something like "no, I put no thought into it whatsoever, the camera was hanging off my bag and the shutter button was pressed accidentally". Like, if a human purposefully took a photo, then they have made choices about location, subject, etc. which have some element of "creativity" to them.
[0] https://www.copyright.gov/engage/photographers/
https://en.wikipedia.org/wiki/Mannion_v._Coors_Brewing_Co.
I remember there being a somewhat similar case in (I think?) Netherlands a few years ago, but currently can't find it.
I know this isn't the case you are referring to, but I am pretty sure cases like this exist more often. This is one of such cases I have in my notes: https://deeplink.rechtspraak.nl/uitspraak?id=ECLI:NL:RBGEL:2...
It's a simple and quite recent Dutch case (feel free to use AI to translate it :p), where the courts basically said that the plaintiff did not sufficiently motivate why their photo would be copyrighted, especially in light of very similar photos having been made by other people (4.5).
There's the monkey selfie incident.
Depends on the jurisdiction. In Germany, practically every photo is copyrightable.
Tell me you cant take good photos without telling me you cant take good photos. Photographers have the skill of colour, framing, perspective, and timing.... and if its nature, they also have to carry heavy bags of camera gear along with their hiking gear and bear spray etc and go out and then cpture photos....national geographic literally made its career off photos.....do you think about what orher weird takes you may have and never notice how wrong they are?
A common technique in paint print shops is to print the piece on canvas, then “add” color to it in globs of acrylic that match, making it stand off from the canvas. A very quick a clever trick to recreate multiples of a piece.
Yeah but it’s super obvious when you pick it up so I don’t know if I would consider it lying.
An easily recognizable lie is still a lie.
This is like saying a dwarf planet is still a planet.
> 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. https://www.copyright.gov/newsnet/2025/1060.html
I think that this means that a single prompt alone does not convey copyright. But if you had spent many hours before the prompt fine tuning the model, or much effort after the prompt shaping the result with further prompts, it could be.
I disagree with this approach because I've seen how much creativity and effort some people can put into slowly evolving a single elaborate prompt. AI can be used as another kind of brush. A prompt can be a masterpiece.
I don't think this is the correct interpretation. I think they mean that if you make something without AI and then modify that with AI, that's covered. Likewise, if you start from an AI output and modify it, that's covered.
But the pure output of a generative model cannot be copyrighted, regardless of how complex the prompt is (note that the prompt itself could be copyrighted).
> But the pure output of a generative model cannot be copyrighted, regardless of how complex the prompt is
If that’s how the court interpreted it, then the software industry is hosed, since that’d mean none of the generated code running in production right now is under any sort of copyright or otherwise protection, lol.
I doubt that much software is entirely AI-generated with no human review or testing, it’s probably more like integrating some public domain snippets you found online into your code (which doesn’t invalidate copyright on the rest of it, or the way it’s put together) or having some files auto-generated by a script (like a C header containing a lookup table for a simple mathematical function, the table isn’t copyrightable itself maybe but the software as a whole still is)
Never mind that.
If a deterministic machine transformation from a copyrightable prompt results in an uncopyrightable image, what do you think a compiler is doing to source code?
AI is not specifically not deterministic from the enduser's perspective. they throw randomness into it and hence why an exact prompt wont produce the same exact result.
a compiler on the other hand is generally pretty deterministic. The non determinism that we see in output is usually non determinism (such as generated dates) in the code that it consumes.
If your argument is that compiler output is more deterministic than image generators, how does that help?
because they are just translating code (that everyone agrees is copyrightable) in a deterministic manner into another medium.
I'm not saying AI art should or shouldn't be copyrightable. One can argue the inputs into the AI generator are copyrightable, but if the output isn't deterministic translation of the input, its a different argument.
The original argument was that AI works wouldn't be copyrightable because they are deterministic, i.e. are just an algorithmic transformation lacking in creativity.
that doesn't seem to be the argument, see top comment (As of now) here
"The courts just take issue with him naming his AI system as the sole author and himself as the copyright owner."
you can't claim a non human as the "author" and claim the material is copyrightable.
the "author" (not the AI) was trying to make a legal point/hack and the courts shot him down.
> human review or testing
Review and testing do not confer a copyright the work reviewed or tested
I guess I was thinking more when it involves rewriting large parts of it because they don't work
That would be authorship, no?
It sounds like they might be under the impression that having any AI-generated output in the code even if parts are human authored would invalidate the copyright, which isn’t true
>If that’s how the court interpreted it, then the software industry is hosed, since that’d mean none of the generated code running in production right now is under any sort of copyright or otherwise protection, lol.
Correct, the jurisprudence there hasn't changed.
I'm not sure this is really true, since copyright applies to distribution.
If you have a substantial amount of backend code (as with most SaaS projects) you're never actually distributing the code, and copyright is never at play. Computer generated artifacts are already in this boat and are protected by virtue of being trade secrets not by copyright.
This could maybe be true of shipping javascript to the browser, which presumably is not going to qualify as a trande secret, but I don't think that's where most companies derive value.
The idea that copyright applies solely to distribution is a popular myth, but it has no support in the actual copyright law. The core exclusive rights in copyright are (in the US, 17 USC § 106):
---
(1) to reproduce the copyrighted work in copies or phonorecords;
(2) to prepare derivative works based upon the copyrighted work;
(3) to distribute copies or phonorecords of the copyrighted work to the public by sale or other transfer of ownership, or by rental, lease, or lending;
(4) in the case of literary, musical, dramatic, and choreographic works, pantomimes, and motion pictures and other audiovisual works, to perform the copyrighted work publicly;
(5) in the case of literary, musical, dramatic, and choreographic works, pantomimes, and pictorial, graphic, or sculptural works, including the individual images of a motion picture or other audiovisual work, to display the copyrighted work publicly; and
(6) in the case of sound recordings, to perform the copyrighted work publicly by means of a digital audio transmission.
---
OTOH, distributing copies created in violation of copyright is a good way to cause legally-cognizable harms to the copyright holder that will increase the potential damage award when you are found liable for copyright infringement, and it also makes it much more likely that someone will notice the infringement in the first place. But its not where the law, on its own terms, begins to apply. Doing any of those without permission (unless it falls into one of the exceptions to copyright protection, like fair use) is a violation of copyright.
>I'm not sure this is really true, since copyright applies to distribution.
Do you mean that a work has to be published prior to registration? What do you mean by "copyright applies to distribution"?
The idea of copyright is to prohibit unauthorized use and reproduction, but none of this actually happens with a proprietary software SaaS backend. You don't actually give anybody the code - they connect to the service.
Access to the service is already governed by computer access laws, which don't depend on copyright. And if you never intentionally distributed your code outside of your org, you can call it a trade secret and nobody else has any legitimate right to access it - whether or not it is copyrightable.
There are other things that aren't copyrightable that are trade secrets already. This would be true of any kind of automated data collection for example. You couldn't copyright it but you can call it a trade secret.
And for any of that stuff, if you want to share it and limit distribution, you just have whoever wants access explicitly agree to be bound by contract law.
>The idea of copyright is to prohibit unauthorized use and reproduction, but none of this actually happens with a proprietary software SaaS backend. You don't actually give anybody the code - they connect to the service.
The point isn't that you have to give it to people, but okay?
>Access to the service is already governed by computer access laws, which don't depend on copyright
Yeah, copyright doesn't control everything, and?
>There are other things that aren't copyrightable that are trade secrets already. This would be true of any kind of automated data collection for example. You couldn't copyright it but you can call it a trade secret.
Okay?
>And for any of that stuff, if you want to share it and limit distribution, you just have whoever wants access explicitly agree to be bound by contract law.
Your point being? You're just rambling assumptions about copyright and other things, which don't even track the actual law.
> Your point being? You're just rambling assumptions about copyright and other things, which don't even track the actual law.
I'm replying to the post that claimed:
> If that’s how the court interpreted it, then the software industry is hosed, since that’d mean none of the generated code running in production right now is under any sort of copyright or otherwise protection, lol.
There is in fact "otherwise protection" for the software industry by... not distributing the code. They don't need copyright over the generated code if they vibe code a SaaS backend. Whether there's copyright or not is irrelevant for the business model.
Copyright is the strongest legal protection available. It does not have a state of mind element. Breach of contract is much more complicated and context-dependent.
>There is in fact "otherwise protection" for the software industry by... not distributing the code.
Copyright protects against reverse engineering in some circumstances, for example.
>Whether there's copyright or not is irrelevant for the business model.
Yeah, I'm going to continue to disagree with you as I'm actually a litigator.
> Yeah, I'm going to continue to disagree with you as I'm actually a litigator.
OK, can you explain to me why this is a disaster for a vibe-coded SaaS? Why are computer access and/or contract laws insufficient and why would a vibe-coded backend be a huge risk?
I really don't understand where copyright on the code itself is necessary to protect these business models, and hopefully you can help fill the gaps.
I didn't say it would be a huge risk, I just disagree that any of those features of the law cover what copyright does. They don't. If a trade secret is ever revealed, all protection is lost. Breach of contract is very complex compared to an infringement claim and would have to be negotiated. As a customer, why would I want to indemnify a software supplier? If there's no indemnity, it's not going to get anyone very far. CFAA basically requires that something get hacked so it's not going to cover the vast majority of scenarios...
>I really don't understand where copyright on the code itself is necessary to protect these business models, and hopefully you can help fill the gaps.
Well, did you ever try to understand? It's so exhausting coming to these threads when people are just making assumptions about how the law works without any regards to what actually happens, and then suggesting policy changes in response.
Here's a scenario - disgruntled ex-employee leaks the code. Now it's free for anyone to use because there is nothing you can do to stop anyone from using it because you have no rights in the code since the trade secret is broken. You can sue the employee. They are probably judgment proof, wont have a lot of money anyway, and will still not stop a competitor from spinning up the same exact thing.
Trade secret was your suggestion by the way... So do, you actually know how trade secrets work, or you just making things up??
Listen, I did try to understand, but I don't know what I don't know, and I'm not an expert like you. It's why I'm asking you for help in this thread.
I'm a layman - I read up on the tools that existed and obviously got it wrong, so I really appreciate you filling the gaps.
“I can stop on an ant, and I can stomp on a flower, so look out, elephants”
What constitutes a modification? Here’s a reverse Sorities Paradox situation.
Let’s say I use an AI prompt to generate an image with 24-bit color, and then I manually change the RGB value of a single pixel from (255,255,255) to (254,255,255).
Does that constitute a modification and would then allow the image to be copyrighted? If not, where is the line?
This is how I understood the original decision a while back - that there had to be some additional element of human involvement post-"gen", though to what extent is still a bit unclear to me.
What's the threshold? Can the person just slap an LUT on an SDXL image in Photoshop and call it a day?
> A prompt can be a masterpiece.
So the true Renaissance artists are the Medicis and the RC church?
> how much creativity and effort
So art is art prompting, or is it creativity and effort? If some toddler spends two hours on a drawing, it's a masterpiece?
> AI can be used as another kind of brush.
A simile does not a truth make.
Its interesting you mention a toddler drawing for 2 hours - I can say with certainty that toddler drawing will demonstrate more creativity than this argument and, every iteration of it, I see every single day. Smh
Actual creators understand what creativity is and what creation is - not all creation even is creative, as its really more of a process, than it is a singular output, and there are monotonous, time consuming, meticulously frustrating parts of the process of creation.
If you want to limit yourself to creation without the quality of life enhancements, that exist in this time and space - you do you.
Don't proselytize tho - you are not doing what you think you are.
> If some toddler spends two hours on a drawing, it's a masterpiece?
A work doesn't have to be a masterpiece to obtain a copyright so what difference does it make? The point of copyright isn't to subjectively critique art.
The prompt itself may be copyrighted.
What matter is whether or not its a masterpiece?
The output is mechanically derived from the prompt. There's zero artistry or creativity involved, and despite all the posturing, those are more or less hard requirements for copyright. The prompt may meet the criteria and be copyrighted.
The outputs of a camera 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 photographic image, or a human makes creative arrangements or modifications of the image, but not the mere pointing of a lens and adjustment of setting.
> a prompt can be a masterpiece.
Then copywrite the prompt, that's always been allowed. Should be just as good if that's the true masterpiece.
Or, as I think we all know, it's not. It's merely a commission, the product is the output. Not the prompt.
AI grifters, new 'prompt copywriting' billable service opportunity just dropped.
>>A prompt can be a masterpiece.
I don't think that word means what you think it means.
You have an extremely low bar for calling something a masterpiece.
A prompt can be clever, insightful, unique, and even uniquely productive.
But it is nowhere near the level of decades-deep skill and creative inspiration required to create art anything worthy of the label "masterpiece".
>>AI can be used as another kind of brush
Perhaps that is a valid analogy, but we do not give copyrights to brushes, no matter how much cost or effort was required to make the brush. The brush is not the only tool required to make the art. To continue the analogy, the artist must also select and mount the canvas, mix and color each shade of paint, build up the base layers, and on and on and on...
It doesn't matter if your "brush" is a five hundred billion dollar machine and you spend six months whispering to it to find just the right incantation to generate your file of pixels — SCOTUS is right, you have not make art to which you can claim a copyright.
And the starving student artist in their garage mixing their paints and using the dollar-store brush did make art worthy of a copyright claim.
>> I disagree with this approach because I've seen how much creativity and effort some people can put into slowly evolving a single elaborate prompt. AI can be used as another kind of brush. A prompt can be a masterpiece.
Absolute nonsense. A work of art is made of semantic stratification, experience, thought process. A prompt lacks all that. AI art can be a tool, but this sentence is a good reminder that on average it’s worth shit all.
Don't forget that the human artist spends hours training on art that predates them, and, in my opinion, that training predisposes the artist to unconsciously replicate elements of art they've trained on previously.
Stop treating humans like machines. Stop normalizing this dehumanizing nonsense.
Am I treating humans like machines or machines like they're human?
LLMs were created in our image. Hallucination, confabulation, sycophancy, psychopathy, learning, reasoning, and blackmail are all behaviours in LLMs that were first found in humans. All these behaviors are present in human writing and imagery captured in a training set. So to me, there's no surprise that LLMs exhibit these behaviors.
Do I think LLMs are sentient or sapient? I'm in the probably not camp. We don't have a good test for either, but they do illustrate the resistance to acknowledging any other being or creation as having the same capabilities as Homo sapiens.
Man it’s hard. This transhumanistic bullshit has fried many good brains
> AI can be used as another kind of brush. A prompt can be a masterpiece.
Sorry, but... cringe. If we are calling prompts "masterpieces" now - letting alone the image generated by it - maybe we don't deserve art at all.
[dead]
> AI can be used as another kind of brush. A prompt can be a masterpiece
What a joke. No, AI is not a brush, it is a slop machine that spits out derivatives of the actual masters. If you go back and forth with a human artist about a commission where you keep nitpicking and wanting adjustments, does that make you the artist? No, it makes you the “ideas guy”
> spits out derivatives of the actual masters
A brief history of art in general.
Sure, but at at least it's created by humans.
I have an allegiance to humans. I have no allegiance to a computer program. That would be pathetic.
[flagged]
As someone who has actually studied art - right back at you.
All the most famous paintings are done like a hundred times - certain scenes were done by ALL Masters, as like a rite of passage or proof of arrival - pretty much the whole time between the OG Masters and modern art.
"Who is your favorite Sebastian?" Is a legit question, for example.
And? Have we all already lost any ability for actual consequential thinking? Each one of those Sebastian is the living proof of a human experience, each with its subtleties, each unique. That is why you ask that question. How on Earth can you compare that to a series of shitty pictures done in a few seconds by a stochastic machine?
You clearly only get sloppy from the machine -> hence your entire understanding, you did some googling, found others with the same experience and you took up this position.
Meanwhile, random person, gets the exact same AI that you used to create literal DaVinci'esque, visibly masterpiece inspired - maybe not "masterpiece" but "masterpiece adjacent" - thats apparently, its not perfect art, but it could have been created in a workshop...
You can't do that. Rather, you cant nake the AI do that.
What is the difference between you and the random person with artworks in the style of the old masters? What do we call that gap?
Isn't that gap normally stuff like talent, ability, skill, knowledge?
All arguments made in this vein are just people whining about their personal lack of ability, as if its a machines fault.
Thanks for the ad hominem. Subjective artistic value aside, do you not dispute that prompting AI to create an artistic image is functionally the same as outsourcing to another human? You input instructions and a commission fee; you get back a piece of artistic expression. You did not create the art, someone else did. And for AI output, the machine cannot claim a copyright on that original image.
If you outsource the image creation, then throw it on a t-shirt design, or modify it, or do whatever, you can copyright the modified work that you Han a hand in creating, but you still are not the creator of the original.
> do you not dispute that prompting AI to create an artistic image is functionally the same as outsourcing to another human?
From who's perspective? Fundamentally, in one situation a human is creating something and in the other situation, no human creates anything.
You generally cannot copyright derivative works as if they are your own, fyi.
Let's say the AI prompt is "Make it black and white". Why does taking a photo and making it gray scale in photoshop result in a copyrightable piece of art, but using an AI model makes the resulting output slop? They seem equivalent to me.
I'll distill it down into something you might understand a bit easier. On social media, such as Instagram, or Tiktok, you'll find a bit of a meme going around that shows the difference between an influencer video of a vacation destination and then a follow-up video from someone with their iPhone, often showing overcrowding of tourists, brown water where there was blue; with these videos often with a poor-sounding Recorder being played over them (I forget the song that it's attempting to play).
The difference between the "real" video and the "influencer" video is the artistry from the artist(s) involved. (And yes, top influencers often have a person or a team of people involved)
Is the Jurassic Park theme btw
Thanks! That's one of them :) I had to look it up as it was bothering me. The other is "September" by Earth, Wind & Fire.
It's not about image modifications, it's about creation. Furthermore, half a similarity on a small aspect doesn't undermine the argument.
Because you're unable to understand art, honestly. Photography isn't just "taking a photo and making a grayscale image in Photoshop"; but rather a combination of a couple of different artistic expression styles that involve understanding how to use the tools you have (a camera, the lens, film or a sensor, and lighting) to capture an expression of an event. Technically speaking, a photo of a mountain isn't just a "photo of a mountain" that you would maybe throw into AI slop--but an actual, legitimate photo of a mountain and how it's captured and presented (no matter the post processing done) is an actual artistic expression of the capture of that mountain. Because absolutely nothing, nothing at all will capture actually standing there looking at the mountain. A photo is the best approximation, and sometimes doesn't even have to be, depending on what the artist wants to express with the image.
In short, your inability to understand photography doesn't justify the use of AI slop to prompt "give me a grayscale image of a mountain" and assume that it's the same thing as a human being taking an actual photo. They're not even close to the same thing.
The original comment is asking from a legal perspective in a very specific example, not an emotional one.
They are describing the artistic qualities that something must posess to be actual artwork - which is relevant to copyrighting artworks, and is also that had been discussed, by Artists and Creators - the entire time
They did not write an emotional comment, they were trying to teach you, bc they ascertained that you don't understand art.
They were right.
>use of AI slop to prompt "give me a grayscale image of a mountain"
That is not the prompt I included in my post. The prompt I gave was for taking an the same photo you would be putting into photoshop and using AI to apply grayscale.
Wait till you learn about comparator mirrors. And renaissance artist studios.
Initialize an algorithm to point your camera at the street and write those bytes to disk and you are the author of a perpetual stream of data.
Initialize an algorithm to point your camera at the street and describe those bytes in words and you are no longer the author a perpetual stream of data.
In a world where slapping an overlay of someone looking incredulous over someone else's video is considered "adding substantive commentary" by every major video sharing platform, I don't even try to understand copyright law at all. It is way over my head.
If it makes seven figures of revenue, there is a real system in place to litigate copyright disputes between corporations. Two kaijiu summoned by ritual magic to fight for the future of the franchise / giant pile of money.
Everything else in the entire system is just bits of monster and building falling randomly. We know if we put the whole population under strict scrutiny ("laser eyes" + "lightning wings"), it would kill every last one of them; every teenager is theoretically criminally liable for the GDP of the Milky Way, a series of violations beginning with a performance of The Birthday Song at their first cake day. Even hiring the cheapest defense lawyer would bankrupt nearly any family in the nation. So we try imperfectly to dodge copyright, hopefully by a couple zip codes, and live in a state of nature on the ground.
> slapping an overlay of someone looking incredulous over someone else's video is considered
it really isn't, you actually have to provide enough relevant commentary for it to be transformative
it just looks like that because
- not every claim leads to a take down, more common is that the advertisement revenue is redirected to the owner of the original video. That is very very common, especially on YT, but not really visible as viewer.
- there are enough copyright holders which overall tolerate reactions, even if they don't fall under fair use.
- Sometimes people claim it doesn't fall under fair use when they don't like how the reaction is done, but that doesn't mean it wouldn't be ruled fair uses if it came in front of court.
- Sometimes people reacting have explicit permission from the original author to do so, no matter if it counts as fair use or not.
and maybe most relevant here, pretty much all large platforms have a tendency to favor the person claiming the copyright violation over the person which reacted to it. To a point there is is sometimes a big problem if systematically abused with false claims.
Those two hypothetical scenarios you listed don’t necessarily work the way you are describing it, which is why the whole logic and mechanisms behind the US copyright laws might seem incomprehensible or illogical to you.
In reality, it is way more complex and less clear-cut. Which makes sense, because oversimplifying it will lead to silly-sounding conclusions and an almost entirely incorrect understanding of how this works.
For those who don’t want to read the actual full explanation (which is a totally normal position, as the explanation is going fairly into the weeds), I will just a put a TLDR summary at the end. I suggest everyone to check out that summary first, and then come back here if there is interest in a more detailed explanation.
----------------------------
First, we gotta settle on 3 key concepts (among many) the US copyright law relies on.
1. Human authorship - self-explanatory; you cannot assign authorship to a fish or your smartphone.
2. Original/minimal creativity - some creative choices, not just "I pressed the button."
3. Fixation - the content needs to be recorded on a tangible medium; you cannot copyright a "mood" or a thought, since those aren’t tangible media.
Now onto your hypothetical scenarios:
1) "Initialize an algorithm to point your camera at the street and write those bytes to disk and you are the author of a perpetual stream of data."
Writing bytes to disk satisfies fixation, but it doesn’t automatically make you the author of a copyrightable work. You gotta satisfy the minimum creativity requirement too (e.g., camera positioning, setup, any other creative choices/actions, etc.). Otherwise you are just running a fully automated security cam feed with zero human input, and those videos aren’t easily copyrightable (if at all). You might own copyright in a video work if there’s sufficient human creative authorship - but mere automated recording doesn’t guarantee that.
2) "Initialize an algorithm to point your camera at the street and describe those bytes in words and you are no longer the author a perpetual stream of data."
This is just close to being plainly incorrect. If you (a human) write a textual description, that text is typically copyrightable as a literary work (assuming it’s not purely mechanical like "frame 1: car, frame 2: another car, etc." with no expressive choices). Creating a description doesn’t erase any copyright you may or may not have had in the underlying recording. They’re just different works (audiovisual work vs. text work).
Important to note: neither makes you the author or owner of the underlying "data" of reality, because copyright protects expression, not the underlying facts.
----------------------------
TLDR:
* Recording the street can produce a copyrightable work if there is human authorship and minimal creativity in how the recording is made. Pure automated capture may fail that.
* Describing the street in words is usually a separate, independently copyrightable work (e.g., a text or audio version of those words), but it doesn’t change the status of the underlying recording.
But how does that apply to photography vs AI photo generation?
Photo (w/ camera): 1. MET: Human authorship - somebody picked the tools (lens, body) and used them.
2. MET: Creativity - somebody chose a subject, lighting, etc.
3. MET: Fixation - film (or SD card)
Photo (w/ AI): 1. MET: Human authorship - somebody picked the tools (models etc) and used them.
2. MET, maybe?: Creativity - somebody wrote the prompt, provided inputs, etc. (how is this substantially different than my wife taking a random snapshot on her phone?)
3. MET: Written to disk, same as a digital camera.
The camera analogy breaks at one specific point: who determines the expressive elements of the final work.
With photography, the human determines framing, angle, timing, lens, exposure. The camera just records light from a scene the human selected and composed. Even a random photo reflects where the photographer stood and when they pressed the shutter. The device doesn’t invent the composition.
With AI imagen, the user provides high-level instructions, but the system determines the actual composition, lighting, geometry, textures, etc. The expressive details of the final image are generated by the model, not directly controlled by the user.
That’s why the US copyright laws currently treat them differently. It is less of a "tool vs. tool", and more of whether the human determined the expressive content (or if the system did). Prompting can be creative (in a legal sense), but giving instructions is not the same as controlling the expression.
If I tell a human painter “paint XYZ in an expressionist style,” I don’t become the author of the painting. The painter does, because they determined the expression. And since the painter (in the case of AI imagen) is not a human, then that work usually cannot be copyrighted.
There is an important caveat to all of this: it’s not binary or perfectly clear-cut. If someone iteratively refines prompts, controls seeds, manually inpaints, selects and arranges outputs, heavily edits the result, etc., then those human contributions can be protected. But purely AI-generated output, where the system determines the expressive elements, is not considered human-authored under the current US copyright laws.
Mind you, none of this is perfectly settled, as this is a very rapidly evolving/adapting area of law (as it pertains to AI usage). I am not claiming that this is the end-all of how it should be legislated or that there are no ways to improve it. But the current reasoning within the US copyright law used to address this type of a scenario (at the present moment) doesn't strike me as illogical or unreasonable.
AI-generated art can't be copyrighted, fine. But what does this mean for the huge spectrum between "I did some fingerpainting" and "Nano Banana spat out this painting"?
What if I use Photoshop and context-aware fill a cloud in? Is that AI-generated or human-generated art?
They cover this. It has to have "substantial human authorship".
So if you start with something you truly made, it would be difficult to use so much context aware fill to negate that.
If you start with something AI generated,at what point does it become copywritable? This is less clear.
But that's fine, because the decision does not torpedo anyone's existing Photoshop workflows.
most likely counts as AI-assisted art, which is copyrightable with you as the owner
like most things copyright there is a gray area there
but in most cases it's either pretty clear and courts would most likely rule in your (copyright holder) if you somehow manage to hit the perfect middle of the gray area
through if you tell the court "the author is my AI" (like in this case), the outcome is pretty obvious
also for better understanding using AI doesn't erase copyright, it just doesn't add it. So if you image was copyrightable before you used an AI tool to change it will stay copyrightable (as long as the original image is still in there to a reasonable degree).
I wonder if Nano Banana spits out an image and I copy it by hand into a different medium like acrylics, chalk, or charcoal. Does a manual transcription suddenly render the image worthy of copyright?
This also raises the meta question: how much does an image need to change to acquire a new copyright? For example, if you change the Last Supper to include two fat Jesuses on either side of the single skinny Jesus, is that enough?
A human couldn't recreate it perfectly in a different medium, thus you are adding human creativity to it, thus copyrightable.
A photograph or a tracing made with an optical device is a mechanical reproduction of the original work, but it has its own copyright because there's a human in the loop.
A forger takes skill and creativity to make a copy of an existing work that is so good it cannot be distinguished from the original. By your assertion, the forgery should have its own independent copyright.
Why wouldn’t that be the case? Each performance of a song is uniquely copyrightable.
I feel like the more important question here is whether AI-generated code can be copyrighted.
Companies responsible for several billion dollars worth of software written over the past ~36 months would really like to know the answer to that one.
For many, many reasons, I suspect patents will become much more important now. After all, it's the ideas that matter now. Which I maintain, has always been the case, because "execution" is nothing more than a series of smaller ideas, except those typically needed money. How convenient for those with capital!
Patents have the drawback of being expensive and very slow to acquire, but having worked on a bunch, they are uniquely suited to be radically optimized by GenAI.
Also patents are very flawed in practice, but the only real protection that is left. Copyright is meaningless when, as people have done, you can reproduce entire saas products by feeding AI screenshots.
Intellectual Property as a whole has been in need for a revamp for a while now, but it's even more critical in the age of AI.
Patents are what allowed the industrial revolution to happen. No one is bringing a cotton gin from idea to design to manufacturing to market if the second you release it every manufacturing company in the world can start making their won.
I wonder how things were invented before patents.
The king paid for it
This strawman pops up often. The purpose of IP is to "promote" progress. The word was carefully chosen.
And luckily we have years of research, empirical and otherwise, showing the impact of IP on innovation and the arts. It's a very complex, multi-dimensional topic where outcomes vary depending on a combination of things like industry, subject matter, time frame, level of economic development of the country, strictness of enforcement, and more.
However overall the impact of IP is much more positive than HN would think, largely because the material and discourse that has gotten airtime here is something that aligns more with the "information wants to be free" crowd.
Patents aren't going to cover the vast majority of what exists as code because patents only cover inventions and not common every day code.
> Copyright is meaningless when, as people have done, you can reproduce entire saas products by feeding AI screenshots.
What do you mean?
> What do you mean?
People are taking screenshots of existing products and feeding them to AI to reproduce them from scratch. From this site: https://ghuntley.com/real/
> Any product features or platforms that were designed for humans. I know that's going to sound really wild, but understand these days I go window-shopping on SaaS companies' websites for product features, rip a screenshot into Claude Code, and it rebuilds that product feature/platform.
Also this comment where backend code could be reproduced wholesale from the APIs: https://news.ycombinator.com/item?id=47259485
And this whole thread about relicensing AI-rewritten open source projects: https://news.ycombinator.com/item?id=47257803
> Patents aren't going to cover the vast majority of what exists as code because patents only cover inventions and not common every day code.
Agreed, but as we're finding out, everyday code is now cheap enough that there will be questions about how much of it is worth protecting... and, given the unclear stance on AI-generated code, maybe even if it is protectable at all.
Yes, the bar for patents is way too high for every day code, but you can always get one by making the claims narrow enough. I think that will actually serve one of the underlying goals of patents by incentivizing people to build things that are actually novel and non-obvious rather than just a slight variation of another project.
For projects where there is absolutely nothing technically novel (which is rarely true, cf "claims narrow enough") moats could be in hoarding data, network effects, and the like.
Note, I'm not necessarily happy about all this, this is just how I think it could play out absent larger changes to IP laws.
>Also this comment where backend code could be reproduced wholesale from the APIs: https://news.ycombinator.com/item?id=47259485
Do you mean deduced instead of reproduced?
>Agreed, but as we're finding out, everyday code is now cheap enough that there will be questions about how much of it is worth protecting... and, given the unclear stance on AI-generated code, maybe even if it is protectable at all.
AI generated code is likely not protectable at all under copyright, but that was expected by anyone who follows this area of the law seriously.
> Yes, the bar for patents is way too high for every day code, but you can always get one by making the claims narrow enough.
Not really, and if you did it would likely not be enforceable because the getting past a §101 motion isn't likely.
>Note, I'm not necessarily happy about all this, this is just how I think it could play out absent larger changes to IP laws.
I'm not sure any of it is a bad thing. Why should AI generated code be protectable under copyright? It can still be protected by trade secret, which seems fairly sufficient for what it ultimately is.
Yes "deduced" makes more sense, though I guess the code was not really the point. I really should have said the backend functionality is being reproduced. As copyright doesn't protect functionality, the closest proxy we had was the code itself, and now that is meaningless.
> AI generated code is likely not protectable at all under copyright, but that was expected by anyone who follows this area of the law seriously.
I haven't followed too closely, but the ruling on AI art seems to leave the door open for "significant human creative input" like "editing, arrangement." This leaves room for photography, but can also be covered by "architecture, followed by iterative prompting and potentially manual editing to refine the code", which is still needed for any non-trivial AI-generated code.
I think the situation is not great because the law is too fuzzy and outdated, and could allow for "AI laundering" of copyrighted works, as the relicensing thread suggests, which is especially not great for open source where trade secrets are not an option.
> Not really, and if you did it would likely not be enforceable because the getting past a §101 motion isn't likely.
Not a lawyer, but I've seen tons of such issued patents out there, and many have even been upheld in lawsuits. §101 is often trivially bypassed adding "computer readable media" or some "non-abstract side-effect" language to claims. Like, one pattern of claims is "A method of - algorithm - algorithm - algorithm - show something to the user / write to a DB / transmit a message / etc."
Perhaps. Even that may not be important if the METR progress line continues much longer, because then all those billion dollars "worth" of software written over the past 3-ish years get re-invented for cents on the dollar.
Separately, I think code is more like an invention than a work of art, and should have been subject only to patent laws instead of (and not in addition to!) copyright laws. This doesn't really make much difference now, as AI doesn't (at least in the UK) have personhood for either copyright or patent law: https://www.briffa.com/blog/can-you-obtain-a-patent-for-inve...
I would assume that the same “substantial human authorship” criterion applies. Copyright is about human creativity, it doesn’t otherwise matter if something is art, prose, typesetting, or code.
Yep. Not just copyright but patents as well
> The Supreme Court previously rejected Thaler's request to hear his argument in a separate case involving prototypes for a beverage holder and a light beacon concerning whether AI-generated inventions should be eligible for U.S. patent protection. His patent applications were rejected by the U.S. Patent and Trademark Office on similar grounds.
https://www.reuters.com/legal/government/us-supreme-court-de...
It doesn't really matter as long as you keep physical control of the code and don't let others copy it.
That would effectively rely on the doctrine of trade secret rather than copyright. A major difference is that accidental or malicious disclosure of a trade secret usually ends the trade secret status, forever. In an alternate universe where computer source code had never been copyrightable, famous leaks (Microsoft Windows, 2004; id Quake, 1997) would have effectively open-sourced those codebases, and other companies could have openly and legally used them.
As source code becomes more of a generated artifact of software development the way object code is an artifact of compilation, we might be moving toward a world where secrecy, constant forward motion, and moats become even more of an asset (vs plain IP protection).
> would have effectively open-sourced those codebases, and other companies could have openly and legally used them.
It's actually better if we keep re-creating the wheel. Keeps more people employed.
Nor does it matter if code has no value.
I do think what happens in this case is SCOTUS will ultimately rule that AI-built code is copyrightable while art is not. I'm sure there's some rationale thick enough for them.
It's strange how hard it is to think of a situation that could lead to that case. Who would bother filing an infringement lawsuit for code whose very existence proves that it can be derived by anyone from LLM prompts? What would the damages even be?
Interesting world we live in. Soon it'll be faster to one-shot the tiny slice of functionality I need from Adobe CS than to navigate their subscription cancellation obstacle course.
> Soon it'll be faster to one-shot the tiny slice of functionality I need from Adobe CS than to navigate their subscription cancellation obstacle course.
Pretty sure you're already in that world. ;)
This is precisely why copyright is practically obsolete. You can't legally forbid someone from paraphrasing, and now we can easily automate it to just within the threshold set by legal cases.
I view this as a massive personal challenge. Can I write instructional materials that are better than an AI summary? I'm going to keep trying, even as books become obsolete.
So I can reverse engineer in peace without Nintendo ninjas lawyers coming after me?
You can always distribute a prompt instead of the product.
Can you imagine the chaos if suddenly all the slop code wasn’t owned by the company? Even though that result would be consistent with this ruling, it undermines the narrative the economy is now riding on, so there will likely be special exemption.
I don't think this article's analysis is accurate. The "human authorship" in Thaler's case wasn't about the abstract concept of human authorship; he literally did not put his name in the "authorship" field of the form, and insisted on review that his name doesn't belong there because he's not the author.
So the ruling doesn't necessarily endorse the Copyright Office's analysis referenced in the article (https://www.copyright.gov/ai/Copyright-and-Artificial-Intell...), and I think that analysis is just not correct. They describe a creator of AI art as simply "re-rolling the dice" when they try different prompts, but that's not correct, clever prompt engineering definitely allows you to "constrain or channel the program’s processing of the sourcenmaterial" and "alter[] the degree of control over the process"
Long standing well known issue, no copyright (in many countries) and (in some countries) non patentable, too.
Through this isn't true for AI assisted art.
And the gray area is very wide and very legal unclear (gray area between human art with AI assistance (e.g. "AI"/transformer architecture based line smoothing or color calibration) and AI art with human touch added to it).
Note that this has very little bearing on the real interesting questions of whether and when human authors can copyright works where AI was used as a tool; this case is specifically about attempts by Thaler to apply for copyright listing an AI as author of a work for which he explicitly denied any human authorship.
Exactly 0 of the artists I know “generate” their images outright - AI generations are always part of a pretty typical artistic workflow.
The way I think of it is this: typical art creation starts from a blank canvas and the artist adds layer upon layer of what you want. Eventually something coherent (to the artist at least) pops out.
AI art starts from a canvas which is filled, and the artist changes the filled canvas to meet their perspective. It’s like those projects where people take a vintage painting and add Pokémon to it. Mostly the people I see using AI art are traditional artists who view it as a new medium in their process, very few “generate” and call it a day.
If we're going to allow AI companies to use copyrighted material in training, the absolute least we could do is prevent copyright of the outputted content.
The irony of someone using AI trained on stolen copyrighted art to generate "art" and then wanting to copyright said "art" is hilarious.
Unpopular opinion: AI is just a new medium of art.
It's like the advent of photography after painting. It was dismissed as an art form for a long time:
- https://antique-photography.com/when-was-photography-conside...
- https://medium.com/@aaronhertzmann/how-photography-became-an... (from 2018!)
Similarly, right now AI art is widely dismissed as "just prompts." But having tried many times to generate images via prompts, it's very hard to get what's in my head to show up in the result. I ended up spending much more time editing the images than creating them... but, I could do that with much simpler tools, without learning advanced tools like Photoshop.
In a couple of instances though, the AI has blown me away by generating something that better captured what I wanted to convey! I suspect the trick is in beng very detailed in where I was coming from and the emotions I wanted to engender.
I predict appreciation of AI art will shift to overall imagination, taste, and appreciation of technical nuances noticeable only to those "skilled in the art", such as prompting techniques and the quirks of the model used. I even suspect there will be genres of AI art using weaker models (kind of like photographs with Polaroid cameras.)
Can you use "I believed this was AI generated" as a copyright defense now?
No? Just as you've never been able to use your ignorance as a defense?
The main reason this is getting really complicated is because copyright inherently goes against the flow of how nature works. It's irrational at its core. It requires increasing amounts of maintenance to sustain.
It doesn't even promote the progress of science and useful arts anymore, and in fact when mixed with our current form of capitalism, it hinders it.
does that mean movies with AI generated art can be repackaged and sold by anyone?
Only if they don’t have “substantial human authorship”.
Add a single pixel manually
That probably wouldn’t count as “substantial human authorship”.
You then own the copyright on that pixel.
Reminds me of "forever minus a day"
I wonder by that logic, can AI generated art violate other’s copyright?
AI usage doesn't remove copyright, it just doesn't gain any new copyright by itself
so a AI based transformation of a copyrighted input is as much a potential copyright violation as a non AI based transformation.
It's just that the human transformation can by potentially itself be seen as art, so if you have a license or fair use you now can have copyright on the transformed peace (with some limitations (1)). And if the transformation is done by AI you won't (but the original authors "partial" copyright on the outcome is still there).
(1): Like if you (human) "transform" a peace of art in 1000 different ways each keeping 0.001% of the original you will likely get 1000x copyright. But if you then use this 1000 peaces you have copyright too to regenerate the original you still have full copyright infringement. In general the law doesn't care about your "trickery" trying to bend laws.
[dupe] Earlier: https://news.ycombinator.com/item?id=47223022
What happens with software now?
Software is now disposable.
Every time I open my phone and find myself back on this comment thread, I find new nonsense.
If you are "anti-AI" and you’ve never changed or evolved your argument - I suggest a pause, a step back and a substantial revaluation.
Some of these comments in this thread - have me wondering if they have actually interacted with an AI.
You are not correct on "principle" - this isn't a moral thing, if you have taken an ethical position - its bc you dont have a functional understanding of how to make it function.
If you were functionally interacting with AI, you would have a more substantial postion, with actual criticism that would have value.
I'm reading a lot of sloppy- written by people, about AI slop.
Why is "AI-generated code" not also "AI art"? What makes "AI-generated code" copyrightable then? Nothing! Being that everything will be made using AI in the future, the courts just suicided the copyright system! Or where exactly does art end and code begin? The same applies to documents and designs.
If I take your AI-generated code file and write it as an artsy-looking image, do I get to deny you copyright?
No-one has checked in court of AI-generated code is copyrightable.
Personally I hope it's not. To me, this is the best outcome for AI in general. If we are going to violate everyone's copyright training AIs, then it's only fair you don't get AI protection on the output.
> Why is "AI-generated code" not also "AI art"?
It is, and if you did what the actor in this case did with code (tried to register a copyright on a piece of code disclaiming any human authorship and claiming an AI tool as the sole author) it would also not be copyrightable.
This case does not actually address what people are actually mostly interested in with copyrightability of works made by humans with the assistance of AI tools, where the human running the tool is claiming authorship, because that was not what happened that was being fought over here.
it's about copyrighted content
so yes it applies to fully AI-generated code as much as to "AI art".
like with AI assisted art it doesn't apply to AI assisted coding
and yes if everything is fully AI generate there is no copyright anymore, that is by design!
Copyright is there to protect human creativity/time investment. If there is no creativity/time investment, then there is no reason for copyright to exist either. Having still copyright there would mean moving it from a law to protect creative work to a law to protect the privileged few which can afford to just mass generate "everything" with AI. That isn't just very undesirable, it's kinda plain evil, as it would mean screwing over the majority of humanity.
Naturally as mentioned that only applies to full AI products, not to AI assisted products in which case the "human contribution" and thinks resulting from it still have copyright.