Skip to main content

Indie game storeFree gamesFun gamesHorror games
Game developmentAssetsComics
SalesBundles
Jobs
TagsGame Engines
you also claim that it’s possible to wash off the authorship and copyright of a work by passing it through an algorithm

I do not need to claim this. That's how copyright works. Look up, what a collage is as an example how weird copyright is. And that's why it is pointless to try argue with copyright against ai.

There's a difference between a function that is reversible and one that is not. Zip is reversible. A learning matrix is not.

I am not arguing against or for ai. I am arguing against your argument against ai. Preference for human work is a far more stable and longer lasting argument. And it will also work against ai that would be trained entirely by commissioned work.

Look up, what a collage is as an example how weird copyright is.

Creating a physical collage is not copyright infringement because no copying is involved, even at the “copying into memory” level. But the creator does not hold sole copyright over the result. You can’t use collage to wash off copyright. The creator doesn’t need copyright in order to display their work.

Collage on a computer, you might be able to get away with it by arguing fair use, depending on the jurisdiction and the lawyers involved. It helps that the collage artist is actually applying their own creativity, unlike an algorithm. But the biggest protection for collage artists is that most copyright holders don’t care enough to go after them. Even if the copyright holder finds out about the collage (a big if), suing is expensive, it’s bad PR, and there’s little to gain by suing the average collage artist. Automated copyright enforcement algorithms like on YouTube absolutely will go after collages.

But my primary argument isn’t about what the law currently is, which varies a lot by jurisdiction and is generally outside of my area of expertise. My argument is about what is morally right, and what the law should be.

Creating a physical collage is not copyright infringement because no copying is involved

If you invoke copyright for arguing, you cannot use a version of copyright that only exists in your head and that you wish would be true.

When collage uses existing works, the result is what some copyright scholars call a derivative work. The collage thus has a copyright separate from any copyrights pertaining to the original incorporated works.

I did not cite a collage to argue that ai are collages. As I said, I cited collages as example how weird copyright law is. Using material for training an ai is currently also ruled as derivative.

And about morality. Well, in my opinion, the morally wrong thing here is not the training of ai. We all stand on the shoulders of giants. But rather the replacement of human work by ai output. If you focus on how ai models are made, your position is vulnerable. The copyright angle does not work. And 100% consented models would defeat that position completely. Meanwhile the ai output of such models would still replace human work.

I did not cite a collage to argue that ai are collages.

And indeed they are not. Collages, unlike AI images, contain creative input beyond their source material, which qualifies them for copyright protection separate from their source material.

Using material for training an ai is currently also ruled as derivative.

Is that what you meant to say? Because it completely undermines your previous arguments.

Derivative works are still bound by the terms of the copyright of the source material. If I create a game that incorporates GPL-licensed code, I have created a derivative work, and I am bound to release my game source code under the terms of the GPL. Furthermore, under United States copyright law, copyright ownership in a derivative work attaches only if the derivative work is lawful, because of a license or other “authorization.”.

But AI images are not derivative works because they are not works in the first place. They are merely derivative. In the US, the output of an algorithm not copyrightable separately from the input.

Well, in my opinion, the morally wrong thing here is not the training of ai. We all stand on the shoulders of giants. But rather the replacement of human work by ai output.

Using procedural art instead of of hiring an artist is not immoral. Using photographs instead of paintings isn’t immoral. Respectfully using public domain art isn’t immoral. Making a text-based game with no art at all isn’t immoral. Nobody is obligated to be a patron of the arts. None of these are based on stealing.

Meanwhile the ai output of such models would still replace human work.

You have no idea how much work goes into training an AI before it starts producing useful output, do you? If each AI were trained only on commissioned artwork produced for that AI, the AI companies could employ every artist on Earth and the AIs still wouldn’t be as good as what we have now.

Curiously how your interpretaion of law still is not working to forbid ai or ai training, is it. The conclusion is, that it does not work this way. Copyright is not the way to fight ai. It's a dead end.

Preference for human work is a better angle.

(1 edit)

Curiously how your interpretaion of law still is not working to forbid ai or ai training, is it.

The typical defense of AI art is that it’s not derivative art but wholly original. Which I think is bullshit, and you seem to think is bullshit, and a whole bunch of other people think is bullshit. But we’re not the ones with the big army of lawyers and we don’t have the courts in our pocket, so the AI companies win. For now. Laws can change, and there’s a lot of anti-AI sentiment in the streets.

One confounding factor is that it’s often incredibly hard to prove where a particular AI image came from. It’s like pollution: we know who’s doing it, we know it’s killing people, but it’s often impossible to match up a single death with a single polluter. Which is why we have separate laws for pollution and for murder, and why it makes more sense to go after the training than after the final product.

At other times, finding the source of an AI image isn’t hard at all.

I was talking about the creation of an ai model. The thing an ai uses to create output from a prompt. Training this thing is ruled derivative by courts already.

Your example is of a real person. If that person has significance to the model, than the model would know how that person looks. Creating an image that looks similar to one of the training images is of course possible. What do you think would happen, if you prompt for the Mona Lisa. This is trivial.

The typical defense of AI art is that it’s not derivative art but wholly original.

From what I gather, it's more a discussion what original means and the notion that all art is derivative. It's a trivially true claim. You do not grow up in a vacuum. You experience things. You cannot paint a tree if you have never seen one and do not even know that such a thing exists. You would construct and extrapolate it from known things, like a shrubbery. And that's how it was in old pictures, when people tried painting animals they never saw. We do stand on the shoulders of giants.

The oldest art you can find is probably cave paintings. What did people paint? Things they saw. How is that "original"?

But this is a philosophical discussion.

Of more practical implication is, that arguing against ai with copyright is fruitless. There are a lot of citations of law suits opened. But what's the outcome so far? The ruling that it is derivative. So accept that and move on to better arguments, if you need justification to avoid ai.

I do not need justifcation, because I made it a matter of taste. I do not like ai in general and prefer human work. I do not need any legal reasons that are already failing. Or arguments about energy, where I personally tried how false these claims are. But this also allows me to appreciate the occasional indie game with ai usage. Just like I occassional play a platfomer, although I dislike the genre.