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.
> You may not use the Service to try to violate the intellectual property rights of others, including copyright, patent, or trademark rights. Doing so may subject you to penalties including legal action or a permanent ban from the Service.
Perplexity's
> Intellectual Property Rights
> Perplexity AI acknowledges and respects the intellectual property rights of all individuals and entities, and expects all users of the Service to do the same. As a user of the Service, you are granted access for your own personal, non-commercial use only.
Yeah, that's nice and all, but it's not what we're talking about. These passages are about deliberately using the tool to violate copyright. What if, in good faith, I don't deliberately attempt to infringe, but the tool still produces results that do? Because that is happening.
And that's just their interpretation of the tool. There is another interpretation that their tool itself is a violation.
I should have explained that those bits are all there is.
You are right, how am I to know that is an image from a movie or passage from the NYT?
Ask George Harrison about "My Sweet Lord" which cost him $587,000 for his unconcious infringement.
Another example would be the 2013 hit "Blurred Lines" by Robin Thicke and Pharrell Williams. It was found to have copied the "feel" and "sound" of Marvin Gaye's 1977 song "Got to Give It Up." The court awarded Gaye's estate $7.4 million in damages, later reduced to $5.3 million.
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.