Because you’re reading this, I’m going to give myself the liberty of presuming that you have seen AND read the Forbes article, “AI-Created Images Aren’t Protected By Copyright Law According To U.S. Copyright Office” and perhaps noticed the following.
“But the drawings, which were all created by Midjourney…”
“the images in the Work […] are not the product of human authorship”
“animals have taken photos…”
“A person who provides text prompts to Midjourney does not ‘actually form’ the generated images […] The information in the prompt may ‘influence’ generated image, but prompt text does not dictate a specific result
And in a much more extensive quote, which by the way appears in the copyrighted Forbes article verbatim,
“there are going to be a lot of lawsuits in the coming months and years as
creators figure out
who owns the intellectual property
created by a machine.
As far as the U.S. Copyright Office is considered, no one does.”
[Image generated by Malcolm Ryder with Dall-E]
Let’s sprint through this quickly. I. A “machine”, being by definition non-human, might just as well be an “animal” despite any mechanism by which they can generate something.
The trendy new label “Creators” always refers to people, not to non-human “generators” of anything. Ownership is not granted to non-humans. The #1 issue here is “ownership” NOT authorship. “Rights” in copyright are about ownership, not about copy.
And ownership is a nonsense term without the concept of “Property”. So it is ultra-clear that copyright is about property rights. It isn’t about creation, nor even about originality. It is about the only force in the universe that cares about property: people.
II. Now, regardless of legal ideas, I have written all of the above using things that were formed by someone else before me — alphabets, words, and grammar. I have used those things to compose “expressions that I did not borrow or copy from someone else.
It is extremely possible, regardless of probability, that many of the composed expressions above could be generated by some other OPERATOR than myself. Grammar is the principle “programming” of the generator that I operate for expressing my compositions.
The fact that I use programming to generate my expressions is irrelevant to the significance of them. And the possibility that my text has been renovated by a spell-checker or grammar-checker, or a human Editor, does not at all disqualify me as the originating producer and communicator.
Meanwhile, if my expressions are understood, no one cares how I materially generated them. They instead might care about the ideas I had that directed my decisions.
At this point, only by arguing against my assertion that I have composed something can you not agree with the following:
Composition is the single most discriminating fact of the generative exercize, that separates the “original” expression from any re-iteration of it.
Composition is, literally, structure.
An unprecedented structure is an “original” structure.
The designer of any unpredecented specific structure is the “author”.
Design is a function and practice that, within its scope, generates structural definition both Prescriptively and Restrictively. By deciding what is to be included, excluded, associated, related, and HOW for all of that, design FORMS the generation regardless of the means that are utilized.
To push this further,
most movies today could not be copyrighted if the requirement was that every frame in the movie must be successfully copyrighted.
An A.I. generated image in a film is there only because someone decided to make it (by either putting it there or leaving it there) a part of the design.
If someone else then copies that image and uses it in a different design, the argument about who “owns” the image is absurd artistically, while it is not absurd in the context of asset management.
III. So, let’s not be stupid about A.I. and copyright.
Owners get rewarded by markets. Authors get rewarded by recognition. A.I. is NOT about “creativity”. And copyright is NOT about “authenticity”. If your goal is to directly reward authorship, that is NOT necessarily the same as rewarding ownership. And ownership does not necessarily coincide with authorship.
Finally, it is necessary to avoid being delusional about how images have valuable meaning. Value is contextual. And meaning occurs across the entire spectrum of cognitive capability from the most elemental and literal to the most complex and abstract. Images do not have to cross some imaginary threshhold along that spectrum to suddenly qualify enough to be especially appreciated for authorship, nor for ownership.
Humans are not the exclusive source of meaningful value in imaging. Context is by far the decisive factor. This doesn’t diminish the thrill of witnessing exceptional human capability. It simply acknowledges that there are many ways to attribute value to original expressions that are generated though images.