But it's not like that. Examples of clearly infringing prompts in TFA were as vague as "animated plumber".

Asking your session musician for something "melancholy" and having them pass off Stairway to Heaven as original would be unreasonable.

You, because you released the song and took the royalties? I don’t think every type of art can be compared against each other though, as there have been numerous precedents specifically for music, some for paintings, and some for photography with their own nuances.

I still think people who are concerned that art related copyright will stifle generative AI should fight copyright laws directly. But that’s a harder pill to swallow since it will cause multi-industry wide havoc.

In your example, there are missing details. Who owns the output? The way you've described it, that would typically mean that the guitarist is creating a "work for hire" so the ownership transfers to you, but that's a contract detail that would need to be resolved.

For whoever owns the output also owns the liability of the output. You yourself might separately be able to pursue a claim against the guitarist for breech of contract. In the process of that, it might get discovered that you deliberately instructed the guitarist to copy the work, or they copied despite your instructions not to.

But that doesn't change the fact that the final work is infringing. It just allows you to pursue damages that could potentially offset any damages you're liable for from the infringement.

But this also isn't exactly the same situation as OpenAI. OpenAI isn't an individual creator working on contact for you. Even if their ToS ultimately assigns copyright of output to you, there is a matter of scale involved that I think changes things. It's one thing if your guitarist damages you by doing shoddy work, it's another of the guitarist systematizes and scales their shoddy work to damage large numbers of people. Perhaps that would then become a class action issue.

But none of the images in the article are for commercial use, they're for private use. So it would be akin to copyright laws saying "If you hire a guitar teacher, they can't play or teach you to play any copyrighted songs. All songs must either be their own original creation or in the public domain."

You are sued for infringement if you are the rightsholder. You need an agreement with the guitarist about rights. The default agreement for session musicians is that you pay them in return for their rights.

It’s like a software engineering contractor. The contractor gets paid, the IP of their work is owned by the company.

If you release a media with copyrighted content it is IMO first and foremost your problem. Now if you have some contract with the guitarist that specifies that he produced a sample he has the rights to and sold it to you, but he clearly wasn't truthful, you can maybe pass the liability to him. This is not, however, how people will hse generative models If you use Dall-E you are not paying OpenAI to buy the rights to a piece Dall-E has produced. I see it more akin to hiring a musician to play for you for an hour, or a painter to paint for you. You are paying OpenAI to paint you something, but you I think OpenAI would never enter a contract which states that they are selling you the rights to a work.

Ed Sheeran just won a case like this.

He basically played the four chord song in court, and showed that the prosecutions’s song was “copying” an earlier song.

https://amp.theguardian.com/music/2023/may/04/ed-sheeran-ver...

The guitarist is not publishing the content, you are.

It could be argued ChatGPT is a publisher too.

using this analogy, copyright holders want to sue the guitarist for having listened to "Stairway to Heaven"