“U.S. Copyright Office Rules A.I. Art Can’t Be Copyrighted An image generated through artificial intelligence lacked the “human authorship” necessary for protection”
Others may have already heard of the ruling. I had not, and it was a bit of a surprise for me considering how prolific everyone seems to be moving toward AI content of all types (including presumably books).
The Supreme Court case Google LLC v. Oracle America Inc. which ruled that “Google hadfair use to copy some of Oracle’s code to develop the Android operating system”, and
The U.S. District Court case Hachette v. Internet Archive] which ruled that the “Internet Archives lending of whole digital copies of books during the pandemic violated fair use”. (They don’t do it anymore.)
The question that immediately popped into my head … Was Google able to hire, or have in-house, better attorneys than the Internet Archive $$$ ?
"provides information on various volunteer lawyers for the arts (VLA) groups and legal clinics that are located across the country. These groups and clinics offer a variety of services to qualifying individual creators and small businesses."
Including 5 pages of volunteer lawyers for the arts (VLA) groups for those who are:
If you are an individual creator or small business in need of legal assistance on a copyright issue and cannot afford to hire an attorney, we encourage you to explore the creator assistance directory below to help you find a VLA or clinic that can answer your questions and assist you with your copyright legal issues.
The list includes VLA’s in many States across the U.S. and one in Vancouver, Canada, 2 in NY and several in California. They also offer Our Free Creator Membership.
I’m pretty sure the US Copyright Office ruling has been discussed here on multiple occasions. I don’t want to rehash that conversation, but a key point is that there is likely to be substantial litigation about exactly how much “human authorship” is required to create a copyrightable work. It is very much not a settled question.
They might, but it’s more likely that using excerpts of computer code is different from redistributing entire books. What Google did is more like taking a quote from an existing work and building an entire new chapter around it.
If I’m not mistaken, I seem to remember that at one time computer code was not copyrightable (or was it patents?) either.
As I read the US Copyright Office ruling, I could think of a few areas where perhaps there might be an opening for the necessary “human authorship” for the CO to rule otherwise, and allow for AI Art et al. to be copyrighted. But for now, as far as I am aware of, the CO ruling is the law.
As for any change, the four questions that came to mind for me were:
How long will it take for the CO ruling to change?
Will the change be in any way retroactive?
What will other jurisdictions (EU, UK, Canada, Japan, China, etc.) do?
Will the change require the IP conventions to meet to work out a common language for the change, and how long will all that take?
This is a very big grey area, but IP lawyers are saying that your best bet is to use an AI like Stable Diffusion where you can give it a guide image, which makes the AI product a derivative work of your own original work, and render out individual elements separately, and then combine and paint over in Photoshop or similar. That means that the AI would be creating a derivative work, and then you are creating another derivative work from that.
You need to disclose your use of AI in the copyright form, so that would be the place to make your claim that the work is protectable.
My reading of the US Copyright Office ruling is along the same lines … I always remember reading something somewhere that stated something along the lines of “the moment you touch it, it’s copyrightable”.
So the type of ‘touching’ that comes from providing a ‘guide image’ I would imagine meets the ‘touching’ test. I realize that there is no such legal test. It’s just a metaphor of where I think things may go as creators fulfill the need to have some human input into the creative process.
Often I feel the law just needs some time when there is the creation of a completely new genera to be fleshed out. Precedence is an importance concept in the law.
Much of the law is base on precedence (despite the recent rulings of our country’s highest court). Any decision that, in retrospect, with hindsight, is likely to set a bad precedence may be difficult to subsequently distinguish, assert and reason.
So to avoid making what may become a mistake in deciding a case that will fly in the face of a new, fuller reality, potentially setting a bad precedence, courts may defer.
In this case, my guess is, given just how fast the world of AI is developing, the US Copyright Office ruling effectively deferred until the ‘greatest invention in human history’ has more flesh on those new bones.
In the meantime, for now, as far as I am aware, the CO ruling is the law of the land in the U.S.
It’s also worth noting that in USA there is no bright-line standard for “infringement” or “fair use.” Instead, it’s “I’ll know it when I see it,” which leaves so much up to the judge or jury that virtually all copyright infringement cases are settled rather than adjudicated.
