𝐿𝒶𝓃𝒶 "not yet begun to fight"A.I. musicians are COOKED. I just recorded an ENTIRE SONG using nothing but my instruments and a few microphones. This would have cost upwards of hundreds of dollars in expensive A.I. subscription fees but for me it was completely FREE. Don't get left behind. This is the future!
Biggles 20X6!@Lana also you can copyright your performance! (ai output is public domain, as per recent SCOTUS ruling)
divVerent@divVerent @Lana more detail here
https://www.morganlewis.com/pubs/2026/03/us-supreme-court-declines-to-consider-whether-ai-alone-can-create-copyrighted-works
Basically you have to document the human involvement part and it has to be enough, or it's public domain.
US Supreme Court Declines to Consider Whether AI Alone Can Create Copyrighted Works
The US Supreme Court has declined to consider the copyrightability of artwork generated purely autonomously by artificial intelligence, leaving in place the “human authorship requirement” for copyright protection. In this LawFlash, we break down the case, the statutory and judicial frameworks governing copyright in autonomously created works, the Copyright Office’s stance, and the practical implications for companies navigating copyright in the age of autonomous AI creation.
@Taco_lad @Lana That is a different question though. They just said that AI generated works cannot be copyrighted by the person who operated the AI, unless part of operating the AI itself constitutes copyright (such as when the prompt itself was already sufficiently detailed to be copyrightable).
They did not say that output of AI is generally in the PD. It can still be encumbered by original works that went into the AI's training data and was in part replicated by the AI, and also original works that were used for prompting the AI.
In particular you can't "AI-wash" someone else's work by feeding it to the AI and asking the AI to recreate it.
They did not say that output of AI is generally in the PD. It can still be encumbered by original works that went into the AI's training data and was in part replicated by the AI, and also original works that were used for prompting the AI.
In particular you can't "AI-wash" someone else's work by feeding it to the AI and asking the AI to recreate it.
@divVerent @Lana correct, you can't use AI to infringe on existing copyrights.
@Taco_lad @Lana And the part I don't understand is why a ruling what necessary for that. That was already the pre-existing rule.
If the input of a data transformation (such as compiling a program) was copyrighted, then so is the output, provided the output contains a sufficient amount of the input - and vice versa.
The court has just figured out that it makes sense to keep applying this rule even if the transformation process is an AI.
Note that the AI can most definitely also encumber a work with additional copyright. E.g. it is quite possible for Google to train Gemini's image generator Nano Banana 2 such that Google can claim copyright on any image it generates (in addition to any copyright on the prompt and on training data by other people that went into it) - we just don't know whether or not Google did. The obvious way to do so would be to train it exclusively with works Google holds the copyright to. Or to have a system prompt such detailed that any reasonable person can recognize the style (well, we do know they didn't do that).
As such, my recommendation regarding use of AI generated works is to not raise the copyright question. Do not release any AI generated source under an open source license. Do not ask money for any AI generated works. If you absolutely must use AI generated works commercially, figure out with the provider of the AI model how the copyright situation works - if you e.g. can make a contract so they license the generated works to you, then you can claim good faith and they would be on the hook if copyrighted works were replicated.
As an example, Google specifically does not provide a license of this kind. Instead, Google says at https://policies.google.com/terms/generative-ai/use-policy: "Do not engage in dangerous or illegal activities, or otherwise violate applicable law or regulations. This includes generating or distributing content that: [...] Violates the rights of others, including privacy and intellectual property rights -- for example, using personal data or biometrics without legally-required consent." - so they shift the blame to the user if anyone's copyright was violated, and make the user responsible for any copyright violations in the generated works, even if the prompt never asked to use someone else's work.
If the input of a data transformation (such as compiling a program) was copyrighted, then so is the output, provided the output contains a sufficient amount of the input - and vice versa.
The court has just figured out that it makes sense to keep applying this rule even if the transformation process is an AI.
Note that the AI can most definitely also encumber a work with additional copyright. E.g. it is quite possible for Google to train Gemini's image generator Nano Banana 2 such that Google can claim copyright on any image it generates (in addition to any copyright on the prompt and on training data by other people that went into it) - we just don't know whether or not Google did. The obvious way to do so would be to train it exclusively with works Google holds the copyright to. Or to have a system prompt such detailed that any reasonable person can recognize the style (well, we do know they didn't do that).
As such, my recommendation regarding use of AI generated works is to not raise the copyright question. Do not release any AI generated source under an open source license. Do not ask money for any AI generated works. If you absolutely must use AI generated works commercially, figure out with the provider of the AI model how the copyright situation works - if you e.g. can make a contract so they license the generated works to you, then you can claim good faith and they would be on the hook if copyrighted works were replicated.
As an example, Google specifically does not provide a license of this kind. Instead, Google says at https://policies.google.com/terms/generative-ai/use-policy: "Do not engage in dangerous or illegal activities, or otherwise violate applicable law or regulations. This includes generating or distributing content that: [...] Violates the rights of others, including privacy and intellectual property rights -- for example, using personal data or biometrics without legally-required consent." - so they shift the blame to the user if anyone's copyright was violated, and make the user responsible for any copyright violations in the generated works, even if the prompt never asked to use someone else's work.