(or, geez, here we go again)
I debated not writing this one at all. It’s about AI again. I promise I’ll find other topics as I go forward. Though technically, I can plead that this one is only AI-adjacent.
It’s really about copyright.
If you are like me, while you’ve enjoyed a lot of the thought process of contemplating what the use of AI means to art, you find the topic is draining these days—not so much because of the angst AI can still rouse between the two camps as the fact that the process for deciding the legal situation is becoming more clear, but at the same time is not yet done. I’m reminded of that time back in the Stone Ages when I’d be waiting at a computer screen for what seemed to be decades, waiting for a pixilated JPG to finish downloading. You could see the colors and shapes, but the details weren’t clear at all.
The AI thing is a problem the world is going to eventually solve, but I wish it would get on with things a bit quicker and let us move forward with full JPG clarity.
Alas, that is not where we are.
Except…
Last week, a member of one of the writing communities I frequent dropped a note about a really cool video he’d come across that had been created by AI. It took me a while, but I eventually watched it, and, yes, from the perspective of technology, it was actually quite cool—though it went on too long for my taste. It was in the form of a music video, and to be honest, as a music video, it was boring simply because the person who made it didn’t include any real narrative that I could grab onto. Maybe it was just me. Dunno. It looked cool, but it made my overall point about AI and art pretty clear. Simply putting AI together for the sake of putting AI together is, at best, a sugar high, and in the end, kind of boring.
Anyway, for reasons that might become obvious, I’m not going to post a link here.
The problem I want to focus on is not the tool the creator used. The problem I want to focus on is that the person who posted it—who is a writer and (to be blunt) really should know better—noted that, while it had some obvious likeness issues and whatnot…wasn’t it so cool!
And, again, yes, it was quite cool–but those “likeness issues” he mentioned stemmed from the fact that the creator used hyper-realistic characters from Star Trek—about which the community member appended the comment that he guessed it was Fan Art, so (I guess), what the hell?
Grrr.
As you might tell from my other conversations on the subject, I’m in what appears to be a small pocket of creators who have no real emotional response to the idea behind the use of AI in creative pursuits. Really, I’m fine with it. To state again, I have not used it in anything beyond play-around mode (because I’m curious, and because I want to actually know what I’m talking about). But if another writer decides to use it and can come up with something that is still in their voice and vision and that their readers actually like (because readers are the final arbitrators), good on them. The business aspects of AI and copyright are another issue, but right now, the courts seem fine with that usage, and the business losses that might (or might not) come from the decision are yours to make.
Who am I to tell someone else how to make their art?
I could go further, but this piece is not meant to be a deep dive into the ins and outs of that philosophy.
The problem here is that this video was not just “Fan Art.”
Like all fan art, this video is a derivative work—which means that, yes, it is a violation of copyright. The video also came from a site that has a lot of that kind of thing on it, and many—if not most—videos on that site had 100K or more views. It’s a monetized site, which means the creator is turning these copyrighted characters into what is almost certainly a pretty good flow of currency.
The whole Fan Fiction thing is an interesting discussion, of course, but the core of the matter with Fan Fiction is that they are all derivative works. Assuming the rights holder learns of the infringement, that puts the response completely in the rights holder’s control. Sometimes the creator is cool with it (it’s still infringing, but the rights holder decides not to pursue damages), and sometimes they are not. If they are not, the rights holder will take legal action (and they will win).
For example, in the days before this post, I saw another person post a silly meme that obviously had been created with AI (which doesn’t really matter), but included an image of Bugs Bunny (which does matter). The person who created this Bugs Bunny meme has broken copyright law, and if it can be tracked, which I assume it can be, is now vulnerable to Warner Brothers’ whims as to whether they will have to pay the statutory charge of at least $150K.
Warner Brothers owns the copyright (likeness) to Bugs Bunny.
In this case, most likely, Warner Brothers will not pursue that meme maker. I assume there’s no revenue stream coming from it, so it would be a PR thing only (note that, unlike trademark issues, copyright infringements do not need to be pursued to have the owner retain copyright, so Warner Brothers loses nothing by leaving that form of theft lie). I doubt they will pay their lawyers to find a meme maker, though they certainly could, and if they decided to that meme maker is in for a world of hurting.
But Paramount finds itself in a different situation.
Here’s a site that’s clearly using material they own, and is clearly making some not insignificant money.
What will happen to this site?
I don’t know.
Unless you work the legal desk at Paramount (the owner of Star Trek copyright), neither do you.
But assuming Paramount finds it (and given the viewership of the place, I assume they will), I expect the answer is not “nothing.” If nothing else, I’d guess the operators of the site will receive a very formal-looking take-down notice. If the money is big enough, though, I’d expect that notice to come along with additional stuff.
But who am I?
Just an outsider looking in.
And to be honest, I’m really more interested in the question of…
What will this bring in the future?
It strikes me that the use of copyrighted visual material through AI may well wind up being handled similarly to how YouTube is handling the cases of independent recording artists making their own cover versions of bigger artists’ songs.
Mary Spender, an indie musician, made a very informative video on this practice a while back. If you’re interested in such mechanics, it’s a fun watch.
The whole process is a touch complicated, but in a nutshell, if you or I were to make our own versions of one of our favorite songs, YouTube would scan it upon upload, then—assuming it finds copyrighted material—rather than just shutting it down, they would forward the video to the rights holder (who has registered with them). Then the rights holder has the ability to allow the use (or not), and if our channels are monetized, can tell YouTube how much of the revenue stream they will take for the usage.
If the rights owner approves, YouTube leaves the version up and splits the revenue per the rights holder’s request. Otherwise, we’re out of luck.
Assuming the right holder approves the use, you or I can decide to leave it up, or—if we feel the rights holder is asking too much—take it down.
A quick check says that YouTube manages in the range of six figures of song uploads per day. Given the ubiquitous nature of AI generation, I can see those numbers exceeded. But I assume the system, which seems to be quite elegant, should scale just fine.
So I’m guessing that something like this will be created to parse AI-generated material that violates copyright—though it’s going to take the world a little time to get up to speed (and until then, the environment will have a bit of a Wild West overtone).
Am I right?
Maybe. Maybe not. Maybe something else will come up to solve the problem better.
That’s how things go with progress, right?
But I think there will be something.
Until then, we are all living in a world where we’re all watching the downloading JPG, waiting for it to snap into focus.
In the meantime, though, it would be great if we could see these things for what they are.
Barring a successful appeal, the use of AI to make your own art (whatever that means) is legally acceptable. But using AI to make something that infringes on copyright, however, like all Fan Art, is not cool (unless the artist/rights holder agrees it’s cool), will not ever be cool (unless…), and also exposes the creator to legal problems.
We should be able to tell those cases apart.
And we should choose wisely.
I am a human. Not an AI. You can tell because keep a Patreon page where I talk about writing and being a writer (among other things). In other words, I post a lot of things there before I post them here. I also share occasional work in progress for Patrons only, and give special discounts and sometimes even free books to Patrons at various levels. If you’d like to support me–or just this blog–you can do so by clicking here:
