In reply to a developer on one of the Linux kernel mailing lists, Linux creator Linus Torvalds firmly put a foot down to push back against anti-AI comments.
The GPL license no longer covers the kernel in legal terms.
The uncopyrightability of AI-written code only applies to the actual strings of code generated by an AI, not to the entire project.
A person could ignore the GPL if they only copied the AI-written portions. But, how could they know for sure which lines were AI generated and which were not? A wrong choice would leave them civilly liable for copyright violation and all they stand to gain would be tiny portions of the Linux kernel code which are worthless by themselves.
There’s no reason to steal the AI generated portions and risk a lawsuit, when you can just generate your own code.
Whoever combines LLM generated code with GPL licensed code is required to license the result under the GPL. Can they do that if the LLM code isn’t copyrightable?
The GPL license is a way to take things that can be copyrighted and grant those rights to others with limitations (the limitations being the terms of the license).
However, the AI generated portions of the code are not copyrightable and so they’re considered public domain and remain so no matter what. Once something is public domain it is always public domain, no matter where it is used.
In practice, if someone used wrote a function by hand and then the AI added some error handling code, then that person has a copyright to the lines that they wrote while the lines that the AI wrote are public domain. The entire function can still be licensed under the GPL but the copyright owner would not be able to pursue you for violation of the license if you used the error handling code in another non-GPL work as that portion is public domain.
From a legal standpoint, since the code isn’t annotated as to which parts were human written (protected by copyright/GPL) and which parts were not (and fall under public domain), it would be unwise to attempt to violate the GPL license terms by taking arbitrary parts of the code and releasing them without a GPL license.
The uncopyrightability of AI-written code only applies to the actual strings of code generated by an AI, not to the entire project.
A person could ignore the GPL if they only copied the AI-written portions. But, how could they know for sure which lines were AI generated and which were not? A wrong choice would leave them civilly liable for copyright violation and all they stand to gain would be tiny portions of the Linux kernel code which are worthless by themselves.
There’s no reason to steal the AI generated portions and risk a lawsuit, when you can just generate your own code.
Whoever combines LLM generated code with GPL licensed code is required to license the result under the GPL. Can they do that if the LLM code isn’t copyrightable?
Essentially yes.
The GPL license is a way to take things that can be copyrighted and grant those rights to others with limitations (the limitations being the terms of the license).
However, the AI generated portions of the code are not copyrightable and so they’re considered public domain and remain so no matter what. Once something is public domain it is always public domain, no matter where it is used.
In practice, if someone used wrote a function by hand and then the AI added some error handling code, then that person has a copyright to the lines that they wrote while the lines that the AI wrote are public domain. The entire function can still be licensed under the GPL but the copyright owner would not be able to pursue you for violation of the license if you used the error handling code in another non-GPL work as that portion is public domain.
From a legal standpoint, since the code isn’t annotated as to which parts were human written (protected by copyright/GPL) and which parts were not (and fall under public domain), it would be unwise to attempt to violate the GPL license terms by taking arbitrary parts of the code and releasing them without a GPL license.
The Free Software Foundation, the creators of the GNU GPL, has an FAQ that covers a lot of the details: https://www.gnu.org/licenses/gpl-faq.en.html#CombinePublicDomainWithGPL