AI/LLMs AI art bans are going to ruin small 3rd party creators

[Maxperson]

Maybe in some future case they'll show how involves they were and it wasn't just "prompt and accept what you get", and the courts will acknowledge it being tool use. I suppose we shall see.

I think that once AI can track behind the image what the artists do to it and everyone can see how much time and design change there is, laws will shift to take that into account and there will be rulings that AI artists can control what they create.

 

[Maxperson]

Your country's courts disagree. They have ruled clearly that AI content is not human created as their reasoning for why it cannot be copyrighted. By using an LLM you have not created anything, not just in my opinion, or from a moral/ethical standpoint, but legally speaking.

It wouldn't be the first time the courts have gotten things wrong. Also, the courts tend to be more cautious with new technologies that are entering the legal space. Over time, as the technology progresses and is better understood, the rulings can and I think probably will change.

 

[Morrus]

It wouldn't be the first time the courts have gotten things wrong.

It isn't. Because they're right.

 

[VHawkwinter]

I think that once AI can track behind the image what the artists do to it and everyone can see how much time and design change there is, laws will shift to take that into account and there will be rulings that AI artists can control what they create.

Like the court case where someone proved pinball had been legally misclassified as gambling, and was in fact a game of skill? I suppose we shall see.

 

[Gorgon Zee]

Are you fully ready to accept all the consequences of claiming the ai is the creator of the works it produces, especially how that relates to the ownership of said works?

I never said that. I said the prompt creator was NOT the creator. That does not mean that the AI is the creator. If you have been following the legal news on AI, you will note that that is indeed the current legal thinking.

 

[Gorgon Zee]

Tools and things cannot create. I'm not a part of the creative process. I am the creative process. The tool is just a thing used to get me to my creative vision.

You are excluding the third option, which is that the result is not actually creative, and that it is simply a non-creative derivative. It’s not that the AI has been creative, or that you have. The result is simply not creative.

If you photocopy a piece of art, neither you nor the photocopier have been creative.

 

[zakael19]

It wouldn't be the first time the courts have gotten things wrong. Also, the courts tend to be more cautious with new technologies that are entering the legal space. Over time, as the technology progresses and is better understood, the rulings can and I think probably will change.

The looong well reasoned and technically adept discussion from the US Copyright Office backing up what they consider to be "human created" vs not for evaluation of copyright was linked earlier in this thread. SCOTUS declined to take up a challenge to it. Until / unless new laws are enacted through congress, their reasoning is likely to stand, as the USCO is part of congress.

Their reasoning is very clear:

"

• Copyright protects the original expression in a work created by a human author, even if the work also includes AI-generated material.
• Copyright does not extend to purely AI-generated material, or material where there is insufficient human control over the expressive elements.
• Whether human contributions to AI-generated outputs are sufficient to constitute authorship must be analyzed on a case-by-case basis.
• Based on the functioning of current generally available technology, prompts do not alone provide sufficient control.
• The case has not been made for additional copyright or sui generis protection for AI-generated content."

They do note they will continue to monitor advances in AI technologies to see if the above changes, but since so far LLMs are not fundamentally changing in how they work from the earlier laughable attempts to today's quite refined outputs, nothing so far seems to warrant that.

The USCO's technically and legally informed view is that "writing a prompt and then watching the black box do its thing" is not sufficiently human authored for legal protections. They even discuss the idea of highly detailed prompts (or even working with an intermediate LLM to generate more machine-accurate prompts for a subsequent system).

 

[FrogReaver]

You are excluding the third option, which is that the result is not actually creative, and that it is simply a non-creative derivative. It’s not that the AI has been creative, or that you have. The result is simply not creative.

If you photocopy a piece of art, neither you nor the photocopier have been creative.

Create does not equal creative.

 

[FrogReaver]

I never said that. I said the prompt creator was NOT the creator. That does not mean that the AI is the creator. If you have been following the legal news on AI, you will note that that is indeed the current legal thinking.

I see the issue. That’s not a logically tenable position.

 

[Maxperson]

• Copyright protects the original expression in a work created by a human author, even if the work also includes AI-generated material.

This does not apply to what I'm talking about. The end result is not going to contain enough of the original work to be recognized as the original work.

• Copyright does not extend to purely AI-generated material, or material where there is insufficient human control over the expressive elements.

This does not apply to what I'm talking about, either. I'm talking about a huge amount of human control.

• Whether human contributions to AI-generated outputs are sufficient to constitute authorship must be analyzed on a case-by-case basis.

This applies, but as I said in my first post, I don't think(but could easily be wrong) that currently the AI programs track what the users do well enough to prove the huge amount of human control that I'm talking about.

• Based on the functioning of current generally available technology, prompts do not alone provide sufficient control.

This does not apply, since I'm not talking about prompts alone.

• The case has not been made for additional copyright or sui generis protection for AI-generated content."

Eventually one(and probably many more) will, though.

They do note they will continue to monitor advances in AI technologies to see if the above changes, but since so far LLMs are not fundamentally changing in how they work from the earlier laughable attempts to today's quite refined outputs, nothing so far seems to warrant that.

The USCO's technically and legally informed view is that "writing a prompt and then watching the black box do its thing" is not sufficiently human authored for legal protections. They even discuss the idea of highly detailed prompts (or even working with an intermediate LLM to generate more machine-accurate prompts for a subsequent system).

Again, I'm not talking about writing a prompt and then watching the black box do it's thing.

Eventually someone is going to do what I suggest, AI will track the changes, the case will go through, and the person will get credit for his or her creation.